Philippa Kaur

Trisos, C. H., Merow, C. & Pigot, A. L. The projected timing of abrupt ecological disruption from climate change. Nature 580, 496–501 (2020).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Schwörer, C. et al. Holocene climate, fire and vegetation dynamics at the treeline in the Northwestern Swiss Alps. Veg. Hist. Archaeobot. 23, 479–496 (2014).Article 

Google Scholar 
Steinbauer, M. J. et al. Accelerated increase in plant species richness on mountain summits is linked to warming. Nature 556, 231–234 (2018).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Grabherr, G., Gottfried, M. & Pauli, H. Climate effects on mountain plants. Nature 369, 448 (1994).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Bennett, K. D. & Willis, K. J. Pollen. Tracking Environmental Change Using Lake Sediments (eds Smol, J. P., Birks, H. J. B., Last, W. M., Bradley, R. S. & Alverson, K.) 5–32 (Kluwer Academic Publishers, 2002).Liu, S. et al. Sedimentary ancient DNA reveals a threat of warming-induced alpine habitat loss to Tibetan Plateau plant diversity. Nat. Commun. 12, 2995 (2021).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Rijal, D. P. et al. Sedimentary ancient DNA shows terrestrial plant richness continuously increased over the Holocene in northern Fennoscandia. Sci. Adv. 7, eabf9557 (2021).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Giguet-Covex, C. et al. Long livestock farming history and human landscape shaping revealed by lake sediment DNA. Nat. Commun. 5, 3211 (2014).Article 
ADS 
PubMed 

Google Scholar 
Väre, H., Lampinen, R., Humphries, C. & Williams, P. Taxonomic diversity of vascular plants in the European alpine areas. in Alpine biodiversity in Europe (eds Nagy, L., Grabherr, G., Körner, C. & Thompson, D. B. A.) 133–148 (Springer Berlin Heidelberg, 2003).Theurillat, J.-P. & Guisan, A. Potential impact of climate change on vegetation in the European alps: A Review. Climatic Change 50, 77–109 (2001).Hewitt, G. The genetic legacy of the Quaternary ice ages. Nature 405, 907–913 (2000).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Tribsch, A. & Schönswetter, P. Patterns of endemism and comparative phylogeography confirm palaeo-environmental evidence for Pleistocene refugia in the Eastern Alps. Taxon 52, 477–497 (2003).Article 

Google Scholar 
Rudmann-Maurer, K., Weyand, A., Fischer, M. & Stöcklin, J. The role of landuse and natural determinants for grassland vegetation composition in the Swiss Alps. Basic Appl. Ecol. 9, 494–503 (2008).Article 

Google Scholar 
Walsh, K. et al. A historical ecology of the Ecrins (Southern French Alps): Archaeology and palaeoecology of the Mesolithic to the Medieval period. Quat. Int. 353, 52–73 (2014).Article 

Google Scholar 
Walsh, K. & Giguet-Covex, C. Encyclopedia of the World’s Biomes 555–573 (Elsevier, 2020).Schwörer, C., Henne, P. D. & Tinner, W. A model-data comparison of Holocene timberline changes in the Swiss Alps reveals past and future drivers of mountain forest dynamics. Glob. Chang. Biol. 20, 1512–1526 (2014).Article 
ADS 
PubMed 

Google Scholar 
Henne, P. D. et al. An empirical perspective for understanding climate change impacts in Switzerland. Reg. Environ. Change 18, 1–17 (2017).
Google Scholar 
Niedrist, G., Tasser, E., Lüth, C., Dalla Via, J. & Tappeiner, U. Plant diversity declines with recent land use changes in European Alps. Plant Ecol. 202, 195–210 (2009).Article 

Google Scholar 
Lasanta-Martínez, T., Vicente-Serrano, S. M. & Cuadrat-Prats, J. M. Mountain Mediterranean landscape evolution caused by the abandonment of traditional primary activities: A study of the Spanish Central Pyrenees. Appl. Geogr. 25, 47–65 (2005).Article 

Google Scholar 
Nautiyal, S. & Kaechele, H. Adverse impacts of pasture abandonment in Himalayan protected areas: Testing the efficiency of a Natural Resource Management Plan (NRMP). Environ. Impact Assess. Rev. 27, 109–125 (2007).Article 

Google Scholar 
Karger, D. N., Nobis, M. P. & Normand, S. CHELSA-TraCE21k v1. 0. Downscaled transient temperature and precipitation data since the last glacial maximum. Climate of the Past (2021).Landolt, E. et al. Flora indicativa: Okologische Zeigerwerte und biologische Kennzeichen zur Flora der Schweiz und der Alpen (Haupt, 2010).Heiri, O., Brooks, S. J., Birks, H. J. B. & Lotter, A. F. A 274-lake calibration data-set and inference model for chironomid-based summer air temperature reconstruction in Europe. Quat. Sci. Rev. 30, 3445–3456 (2011).Article 
ADS 

Google Scholar 
Heiri, O., Ilyashuk, B., Millet, L., Samartin, S. & Lotter, A. F. Stacking of discontinuous regional palaeoclimate records: Chironomid-based summer temperatures from the Alpine region. Holocene 25, 137–149 (2015).Article 
ADS 

Google Scholar 
Ivy-Ochs, S. et al. Latest Pleistocene and Holocene glacier variations in the European Alps. Quat. Sci. Rev. 28, 2137–2149 (2009).Article 
ADS 

Google Scholar 
Finsinger, W. & Tinner, W. Pollen and plant macrofossils at Lac de Fully (2135 m a.s.l.): Holocene forest dynamics on a highland plateau in the Valais, Switzerland. Holocene 17, 1119–1127 (2007).Article 
ADS 

Google Scholar 
Baroni, C. et al. Last Lateglacial glacier advance in the Gran Paradiso Group reveals relatively drier climatic conditions established in the Western Alps since at least the Younger Dryas. Quat. Sci. Rev. 255, 106815 (2021).Article 

Google Scholar 
Schibler, J., Elsner, J. & Schlumbaum, A. Incorporation of aurochs into a cattle herd in Neolithic Europe: Single event or breeding? Sci. Rep. 4, 5798 (2014).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Schimmelpfennig, I. et al. A chronology of Holocene and Little Ice Age glacier culminations of the Steingletscher, Central Alps, Switzerland, based on high-sensitivity beryllium-10 moraine dating. Earth Planet. Sci. Lett. 393, 220–230 (2014).Article 
ADS 
CAS 

Google Scholar 
Ilyashuk, E. A., Heiri, O., Ilyashuk, B. P., Koinig, K. A. & Psenner, R. The Little Ice Age signature in a 700-year high-resolution chironomid record of summer temperatures in the Central Eastern Alps. Clim. Dyn. 52, 1–15 (2018).
Google Scholar 
Willerslev, E. et al. Fifty thousand years of Arctic vegetation and megafaunal diet. Nature 506, 47–51 (2014).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Alsos, I. G. et al. Ancient sedimentary DNA shows rapid post-glacial colonisation of Iceland followed by relatively stable vegetation until the Norse settlement (Landnám) AD 870. Quat. Sci. Rev. 259, 106903 (2021).Article 

Google Scholar 
Pansu, J. et al. Reconstructing long-term human impacts on plant communities: an ecological approach based on lake sediment DNA. Mol. Ecol. 24, 1485–1498 (2015).Article 
PubMed 

Google Scholar 
Varotto, C. et al. A pilot study of eDNA metabarcoding to estimate plant biodiversity by an alpine glacier core (Adamello glacier, North Italy). Sci. Rep. 11, 1208 (2021).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Parducci, L. et al. Proxy comparison in ancient peat sediments: Pollen, macrofossil and plant DNA. Philos. Trans. R. Soc. Lond. B Biol. Sci. 370, 20130382 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Clarke, C. L. et al. A 24,000-year ancient DNA and pollen record from the Polar Urals reveals temporal dynamics of arctic and boreal plant communities. Quat. Sci. Rev. 247, 106564 (2020).Article 

Google Scholar 
Niemeyer, B., Epp, L. S., Stoof-Leichsenring, K. R., Pestryakova, L. A. & Herzschuh, U. A comparison of sedimentary DNA and pollen from lake sediments in recording vegetation composition at the Siberian treeline. Mol. Ecol. Resour. 17, e46–e62 (2017).Article 
CAS 
PubMed 

Google Scholar 
Wilson, J. B., Peet, R. K., Dengler, J. & Pärtel, M. Plant species richness: the world records. J. Veg. Sci. 23, 796–802 (2012).Article 

Google Scholar 
Wick, L., van Leeuwen, J. F. N., van der Knaap, W. O. & Lotter, A. F. Holocene vegetation development in the catchment of Sägistalsee (1935 m asl), a small lake in the Swiss Alps. J. Paleolimnol. 30, 261–272 (2003).Article 
ADS 

Google Scholar 
Lotter, A. F. et al. Holocene timber-line dynamics at Bachalpsee, a lake at 2265 m a.s.l. in the northern Swiss Alps. Veg. Hist. Archaeobot. 15, 295–307 (2006).Article 

Google Scholar 
Thöle, L. et al. Reconstruction of Holocene vegetation dynamics at Lac de Bretaye, a high-mountain lake in the Swiss Alps. Holocene 26, 380–396 (2016).Article 
ADS 

Google Scholar 
Heiri, O., Lotter, A. F., Hausmann, S. & Kienast, F. A chironomid-based Holocene summer air temperature reconstruction from the Swiss Alps. Holocene 13, 477–484 (2003).Article 
ADS 

Google Scholar 
Garcés-Pastor, S., Cañellas-Boltà, N., Clavaguera, A., Calero, M. A. & Vegas-Vilarrúbia, T. Vegetation shifts, human impact and peat bog development in Bassa Nera pond (Central Pyrenees) during the last millennium. Holocene 27, 553–565 (2017).Article 
ADS 

Google Scholar 
Aeschimann, D., Lauber, K., Moser, D. M. & Theurillat, J. P. Flora Alpina: Atlas des 4500 Plantes Vasculaires des Alpes (Belin, 2004).Sønstebø, J. H. et al. Using next-generation sequencing for molecular reconstruction of past Arctic vegetation and climate. Mol. Ecol. Resour. 10, 1009–1018 (2010).Article 
PubMed 

Google Scholar 
Diekmann, M. Species indicator values as an important tool in applied plant ecology—a review. Basic Appl. Ecol. 4, 493–506 (2003).Article 

Google Scholar 
Giesecke, T. et al. Postglacial change of the floristic diversity gradient in Europe. Nat. Commun. 10, 5422 (2019).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Colombaroli, D. & Tinner, W. Determining the long-term changes in biodiversity and provisioning services along a transect from Central Europe to the Mediterranean. Holocene 23, 1625–1634 (2013).Article 
ADS 

Google Scholar 
Schwörer, C., Colombaroli, D., Kaltenrieder, P., Rey, F. & Tinner, W. Early human impact (5000–3000 BC) affects mountain forest dynamics in the Alps. J. Ecol. 103, 281–295 (2015).Article 

Google Scholar 
Furtwängler, A. et al. Ancient genomes reveal social and genetic structure of Late Neolithic Switzerland. Nat. Commun. 11, 1915 (2020).Article 
ADS 
PubMed 
PubMed Central 

Google Scholar 
Gilck, F. & Poschlod, P. The origin of alpine farming: A review of archaeological, linguistic and archaeobotanical studies in the Alps. Holocene 29, 1503–1511 (2019).Article 
ADS 

Google Scholar 
Tinner, W., Nielsen, E. H. & Lotter, A. F. Mesolithic agriculture in Switzerland? A critical review of the evidence. Quat. Sci. Rev. 26, 1416–1431 (2007).Article 
ADS 

Google Scholar 
Berthel, N., Schwörer, C. & Tinner, W. Impact of Holocene climate changes on alpine and treeline vegetation at Sanetsch Pass, Bernese Alps, Switzerland. Rev. Palaeobot. Palynol. 174, 91–100 (2012).Article 

Google Scholar 
Hafner, A. & Schwörer, C. Vertical mobility around the high-alpine Schnidejoch Pass. Indications of Neolithic and Bronze Age pastoralism in the Swiss Alps from paleoecological and archaeological sources. Quat. Int. https://doi.org/10.1016/j.quaint.2016.12.049 (2017).Oveisi, M. et al. Potential for endozoochorous seed dispersal by sheep and goats: Risk of weed seed transport via animal faeces. Weed Res. 61, 1–12 (2021).Article 

Google Scholar 
Bardgett, R. D. & Wardle, D. A. Herbivore-mediated linkages between aboveground and belowground communities. Ecology 84, 2258–2268 (2003).Article 

Google Scholar 
Scherrer, D. & Körner, C. Topographically controlled thermal-habitat differentiation buffers alpine plant diversity against climate warming. J. Biogeogr. 38, 406–416 (2011).Article 

Google Scholar 
Giguet-Covex, C. et al. New insights on lake sediment DNA from the catchment: Importance of taphonomic and analytical issues on the record quality. Sci. Rep. 9, 14676 (2019).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Andres, B. Alpine settlement remains in the Bernese Alps (Switzerland) in medieval and modern times. Historical Archaeologies of Transhumance across Europe (eds Costello, E. & Svensson, E.) 155–169 (Routledge, 2018).eTopoi. Journal for Ancient Studies. 3, 279–283 (2012).Grime, J. P. Competitive exclusion in herbaceous vegetation. Nature 242, 344–347 (1973).Article 
ADS 

Google Scholar 
Yuan, Z. Y., Jiao, F., Li, Y. H. & Kallenbach, R. L. Anthropogenic disturbances are key to maintaining the biodiversity of grasslands. Sci. Rep. 6, 22132 (2016).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Spiegelberger, T., Matthies, D., Müller-Schärer, H. & Schaffner, U. Scale-dependent effects of land use on plant species richness of mountain grassland in the European Alps. Ecography 29, 541–548 (2006).Article 

Google Scholar 
Maurer, K., Weyand, A., Fischer, M. & Stöcklin, J. Old cultural traditions, in addition to land use and topography, are shaping plant diversity of grasslands in the Alps. Biol. Conserv. 130, 438–446 (2006).Article 

Google Scholar 
Kampmann, D. et al. Mountain grassland biodiversity: Impact of site conditions versus management type. J. Nat. Conserv. 16, 12–25 (2008).Article 

Google Scholar 
Pellegrini, E., Buccheri, M., Martini, F. & Boscutti, F. Agricultural land use curbs exotic invasion but sustains native plant diversity at intermediate levels. Sci. Rep. 11, 8385 (2021).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Bakker, E. S., Ritchie, M. E., Olff, H., Milchunas, D. G. & Knops, J. M. H. Herbivore impact on grassland plant diversity depends on habitat productivity and herbivore size. Ecol. Lett. 9, 780–788 (2006).Article 
PubMed 

Google Scholar 
Speed, J. D. M., Austrheim, G., Hester, A. J. & Mysterud, A. Elevational advance of alpine plant communities is buffered by herbivory. J. Veg. Sci. 23, 617–625 (2012).Article 

Google Scholar 
Filazzola, A. et al. The effects of livestock grazing on biodiversity are multi-trophic: A meta-analysis. Ecol. Lett. 23, 1298–1309 (2020).Article 
PubMed 

Google Scholar 
Evans, D. M. et al. The cascading impacts of livestock grazing in upland ecosystems: A 10-year experiment. Ecosphere 6, art42 (2015).Article 

Google Scholar 
Alexander, J. M., Diez, J. M. & Levine, J. M. Novel competitors shape species’ responses to climate change. Nature 525, 515–518 (2015).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Mathieu, J. Eine Agrargeschichte der inneren Alpen. Graubünden, Tessin, Wallis 1500–1800 (Chronos, 1992).Aerni, K, Egli, H. R & Fehn, K. Siedlungsprozesse an der Höhengrenze der Ökumene: am Beispiel der Alpen: Referate der 16 Tagung des” Arbeitskreises für genetische Siedlungsforschung in Mitteleuropa” vom 20.−23. (Siedlungsforschung: Spiez, 1991).Brugger, S. O. et al. Alpine glacier reveals ecosystem impacts of Europe’s prosperity and peril over the last millennium. Geophys. Res. Lett. 48, e2021GL095039 (2021).Merkt, J. & Streif, H. Stechrohr-Bohrgeräte für limnische und marine Lockersedimente. Geologisches Jahrbuch 88, 137–148 (1970).Lamb, A. L. Determination of organic and carbonate content in soils and sediments by loss on ignition (LOI), NERC Isotope Geosciences Laboratory Report, 197 (2004).Reimer, P. J. et al. The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon https://doi.org/10.1017/RDC.2020.41 (2020).Blaauw, M. & Christen, J. A. Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Anal. 6, 457–474 (2011).Article 
MathSciNet 
MATH 

Google Scholar 
Brooks, S. J., Langdon, P. G. & Heiri, O. The identification and use of Palaearctic Chironomidae larvae in palaeoecology. Quat. Res. Assoc. i-vi, 1-276 (2007).Schulze, E. A Key to the Larval Chironomidae and their Instars from Austrian Danube Region Streams and Rivers with Particular Reference to a Numerical Taxonomic Approach. Part I. In: Wasser und Abwasser, Supplementband 3/93. Hrsg.: Bundesamt für Wassergüte, Wien-Kaisermühlen. Schriftenleitung: Werner Kohl. Selbstverlag, 1993, 514 S., öS 562. Acta Hydrochim. Hydrobiol. 22, 191–191 (1994).Article 

Google Scholar 
Juggins, S. C2: Software for ecological and palaeoecological data analysis and visualisation (user guide version 1.5). Newcastle upon Tyne: Newcastle University (2007). https://www.staff.ncl.ac.uk/stephen.juggins/software/code/C2.pdf.Moore, P. D., Webb, J. A. & Collison, M. E. Pollen Analysis, edn 2 (Blackwell, 1991).Stockmarr & Ja Tabletes with spores used in absolute pollen analysis. Pollen Spores 13, 615–621 (1971).
Google Scholar 
Reille, M. Pollen et spores d’Europe et d’Afrique du Nord (Laboratoire de Botanique historique et Palynologie, Marseille, 1992).van Geel, B. et al. Environmental reconstruction of a Roman Period settlement site in Uitgeest (The Netherlands), with special reference to coprophilous fungi. J. Archaeol. Sci. 30, 873–883 (2003).Article 

Google Scholar 
Bennett, K. D. Determination of the number of zones in a biostratigraphical sequence. N. Phytol. 132, 155–170 (1996).Article 
CAS 

Google Scholar 
Tinner, W. et al. Pollen and charcoal in lake sediments compared with historically documented forest fires in southern Switzerland since AD 1920. Holocene 8, 31–42 (1998).Article 
ADS 

Google Scholar 
Adolf, C. et al. The sedimentary and remote-sensing reflection of biomass burning in Europe. Glob. Ecol. Biogeogr. 27, 199–212 (2018).Article 

Google Scholar 
Tinner, W. & Hu, F. S. Size parameters, size-class distribution and area-number relationship of microscopic charcoal: Relevance for fire reconstruction. Holocene 13, 499–505 (2003).Article 
ADS 

Google Scholar 
Parducci, L. et al. Ancient plant DNA in lake sediments. N. Phytol. 214, 924–942 (2017).Article 
CAS 

Google Scholar 
Alsos, I. G. et al. The treasure vault can be opened: Large-scale genome skimming works well using herbarium and silica gel dried material. Plants 9, 432 (2020).Article 
CAS 
PubMed Central 

Google Scholar 
Taberlet, P. et al. Power and limitations of the chloroplast trnL (UAA) intron for plant DNA barcoding. Nucleic Acids Res. 35, e14 (2007).Article 
PubMed 

Google Scholar 
Voldstad, L. H. et al. A complete Holocene lake sediment ancient DNA record reveals long-standing high Arctic plant diversity hotspot in northern Svalbard. Quat. Sci. Rev. 234, 106207 (2020).Article 

Google Scholar 
Boyer, F. et al. obitools: A unix-inspired software package for DNA metabarcoding. Mol. Ecol. Resour. 16, 176–182 (2016).Article 
CAS 
PubMed 

Google Scholar 
Ficetola, G. F. et al. An in silico approach for the evaluation of DNA barcodes. BMC Genomics 11, 434 (2010).Article 
PubMed 
PubMed Central 

Google Scholar 
Soininen, E. M. et al. Highly overlapping winter diet in two sympatric lemming species revealed by DNA metabarcoding. PLoS One 10, e0115335 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Boratyn, G. M. et al. BLAST: A more efficient report with usability improvements. Nucleic Acids Res. 41, W29–W33 (2013).Article 
PubMed 
PubMed Central 

Google Scholar 
Leonard, J. A. et al. Animal DNA in PCR reagents plagues ancient DNA research. J. Archaeol. Sci. 34, 1361–1366 (2007).Article 

Google Scholar 
Deiner, K. et al. Environmental DNA metabarcoding: Transforming how we survey animal and plant communities. Mol. Ecol. 26, 5872–5895 (2017).Article 
PubMed 

Google Scholar 
Ter Braak, C. J. F. & Prentice, I. C. A theory of gradient analysis. Adv. Ecol. Res. 18, 271–317 (Elsevier, 1988).Vieira, D. C., Brustolin, M. C., Ferreira, F. C. & Fonseca, G. segRDA: Anr package for performing piecewise redundancy analysis. Methods Ecol. Evol. 10, 2189–2194 (2019).Article 

Google Scholar 
Simpson, G. L. Modelling palaeoecological time series using generalised additive models. Front. Ecol. Evol. 6, 149 (2018).Wood, S. N. Generalized Additive Models: An Introduction with R (Chapman and Hall/CRC, 2017).Lefcheck, J. S. piecewiseSEM: Piecewise structural equation modelling inr for ecology, evolution, and systematics. Methods Ecol. Evol. 7, 573–579 (2016).Article 

Google Scholar 
Chen, W. & Ficetola, G. F. Numerical methods for sedimentary‐ancient‐DNA‐based study on past biodiversity and ecosystem functioning. Environ. DNA 2, 115–129 (2020).Article 

Google Scholar 
Juggins, S. Rioja: Analysis of Quaternary Science Data. R package version 0.9-26. https://cran.r-project.org/web/packages/rioja/index.html (2020).Oksanen, J. et al. vegan: Community Ecology Package. Software http://CRAN.R-project.org/package=vegan (2012).Wickham, H. ggplot2-Elegant Graphics for Data Analysis (Springer, 2016).Fox, J. & Weisberg, S. An R Companion to Applied Regression (Sage, 2019).Tinner, W. & Ammann, B. Long-term responses of mountain ecosystems to environmental changes: Resilience, adjustment, and vulnerability. In Global change and mountain regions. 133–143 (Springer, Dordrecht; 2005). More

Davies EK, Peters AD, Keightley PD (1999) High frequency of cryptic deleterious mutations in Caenorhabditis elegans. Science 285:1748–1751Article 
CAS 
PubMed 

Google Scholar 
Eyre-Walker A, Keightley PD (2007) The distribution of fitness effects of new mutations. Nat Rev Genet 8:610–618Article 
CAS 
PubMed 

Google Scholar 
Eyre-Walker A, Keightley PD (2013) A comparison of models to infer the distribution of fitness effects of new mutations. Genetics 193:1197–1208Article 

Google Scholar 
Fry JD, Keightley PD, Heinsohn SL, Nuzhdi SV (1999) New estimates of the rates and effects of mildly deleterious mutation in Drosophila melanogaster. Proc Natl Acad Sci 96:574–579Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
García-Dorado A (2007) Shortcut predictions for fitness properties at the mutation-selection-drift balance and for its buildup after size reduction under different management strategies. Genetics 176:983–997Article 
PubMed 
PubMed Central 

Google Scholar 
García-Dorado A (2012) Understanding and predicting the fitness decline of shrunk populations: inbreeding, purging, mutation, and standard selection. Genetics 190:1461–1476Article 
PubMed 
PubMed Central 

Google Scholar 
García-Dorado A (2015) On the consequences of ignoring purging on genetic recommendations for minimum viable population rules. Heredity 115:185–187Article 
PubMed 
PubMed Central 

Google Scholar 
García-Dorado A, Caballero A (2021) Neutral genetic diversity as a useful tool for conservation biology. Conserv Genet 22:541–545Article 

Google Scholar 
Garner BA, Hoban S, Luikart G (2020) IUCN Red List and the value of integrating genetics. Conserv Genet 21:795–801Article 

Google Scholar 
Hedrick PW, García-Dorado A (2016) Understanding inbreeding depression, purging, and genetic rescue. Trends Ecol Evol 31:940–952Article 
PubMed 

Google Scholar 
Kardos M, Armstrong EE, Fitzpatrick SW, Hauser S, Hedrick PW, Miller J et al. (2021) The crucial role of genome-wide genetic variation in conservation. Proc Natl Acad Sci USA 118:e2104642118Khan A, Patel A, Shukla H, Viswanathan A, van der Valk T, Borthakur U, … & Ramakrishnan U (2021) Genomic evidence for inbreeding depression and purging of deleterious genetic variation in Indian tigers. Proc. Natl. Acad. Sci. 118Kimura M, Maruyama T, Crow JF (1963) The mutation load in small populations. Genetics 48:1303–1312Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Kimura M (1980) Average time until fixation of a mutant allele in a finite population under continued mutation pressure: Studies by analytical, numerical, and pseudo-sampling methods. Proc Natl Acad Sci 77:522–526Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Morin PA, Archer FI, Avila CD, Balacco JR, Bukhman YV, Chow, W, … & Jarvis ED (2021) Reference genome and demographic history of the most endangered marine mammal, the vaquita. Mol Ecol Resour 21:1008–1020Mukai T (1964) The genetic structure of natural populations of Drosophila melanogaster. I. Spontaneous mutation rate of polygenes controlling viability. Genetics 50:1–19Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Nietlisbach P, Muff S, Reid JM, Whitlock MC, Keller LF (2019) Nonequivalent lethal equivalents: Models and inbreeding metrics for unbiased estimation of inbreeding load. Evol Applic 12:266–279Article 

Google Scholar 
O’Grady JJ, Brook BW, Reed DH, Ballou JD, Tonkyn DW, Frankham R (2006) Realistic levels of inbreeding depression strongly affect extinction risk in wild populations. Biol Conserv 133:42–51Article 

Google Scholar 
Pérez-Pereira N, Caballero A, García-Dorado A (2021) Reviewing the consequences of genetic purging on the success of rescue programs. Conserv Gen 23:1–17Article 

Google Scholar 
Pérez-Pereira N, Wang J, Quesada H, Caballero A (2022). Prediction of the minimum effective size of a population viable in the long term. Biodivers Conserv https://doi.org/10.1007/s10531-022-02456-zRobinson JA, Kyriazis CC, Nigenda-Morales SF, Beichman AC, Rojas-Bracho L, Robertson KM et al. (2022) The critically endangered vaquita is not doomed to extinction by inbreeding depression. Science 376:635–639Article 
CAS 
PubMed 

Google Scholar 
Teixeira JC, Huber CD (2021) The inflated significance of neutral genetic diversity in conservation genetics. Proc Natl Acad Sci USA 118:e2015096118Wade EE, Kyriazis C, Cavassim MIA, Lohmueller KE (2022) Quantifying the fraction of new mutations that are recessive lethal. bioRxiv 1–24, https://www.biorxiv.org/content/10.1101/2022.04.22.489225v1 More

The framework of Swin MSIWe established the first Swin Transformer-based mosquito species identification (Swin MSI) model, with the help of self-constructed image dataset and multi-adjustment-test. Gradient-weighted class activation mapping was used to visualize the identification process (Fig. 1a). The key Swin Transformer block was described on Fig. 1b. Based on practical needs, Swin MSI was additional designed to identify Culex pipiens Complex on the subspecies level (Fig. 1c) and novel mosquito (which was defined as ones beyond 17 species in our dataset) classification attribution (Fig. 1d). Detailed results are shown in the following sections.Figure 1The Framework of Swin MSI. (a)The basic architecture for mosquito features extraction and identification. Attention visualization generated by filters at each layer are shown. (b) Details for Swin Transformer block. (c) For mosquito within our dataset 17 species, output is the top 5 confidence species. (d) For mosquito beyond 17 species (defined as novel species), whether the output is a species or a genus is decided after comparing with confidence threshold.Full size imageMosquito datasetsWe established the highest-definition and most-balanced mosquito image dataset to date. The mosquito image dataset covers 7 genera and 17 species (including 3 morphologically similar subspecies in the Cx. pipiens Complex), which covers the most common and important disease-transmitting mosquitoes at the global scale, with a total of 9,900 mosquito images. The image resolution was 4464 × 2976 pixels. The specific taxonomic status and corresponding images are shown in Fig. 2. Due to the limitation of field collection, Ae. vexans, Coquillettidia ochracea, Mansonia uniformis, An. vagus and Toxorhynchites splendens only have females or only have males. In addition, each mosquito species included 300 images of both sexes, which was large enough and same number for each species, in order to balance the capacity and variety of training sets.Figure 2Taxonomic status and index of mosquito species included in this study Both male and female mosquitoes were photographed from different angles such as dorsal, left side, right side, ventral side, etc. Except for 5 species, each mosquito includes 300 images of both sexes, and the resolution of mosquito photos were 4464 × 2976. Cx. pipiens quinquefasciatus, Cx. pipiens pallens, and Cx. pipiens molestus (subspecies level, in dark gray background) were 3 subspecies in Cx. pipiens Complex (species level).Full size imageWorkflow for mosquito species identificationA three-stage flowchart of building best deep learning model for identification of mosquito species model was adopted (Fig. 3). The first learning stage was conducted by three CNNs (the Mask R-CNN, DenseNet, and YOLOv5) and three transformer models (the Detection Transformer, Vision Transformer, and Swin Transformer). Based on the performance of the first-stage model and the real mosquito labels, the second learning stage involved adjusting the model parameters of the three Swin Transformer variants (T, B, and L) to compare their performances. The third learning stage involved testing the effects of inputting differently sized images (384 × 384 and 224 × 224) to the Swin Transformer-L model; finally, we proposed a deep learning model for mosquito species identification (Swin MSI) to test the recognition effects of different mosquito species. The model was validated on different mosquito species, with a focus on the identification accuracy of three subspecies within the Cx. pipiens Complex and the detection effect of novel mosquito species.Figure 3Flowchart of testing deep learning model for intelligent identification of mosquito species.Full size imageComparison between the CNN model and Transformer model results (1st round of learning)Figure 4a shows the accuracies obtained for the six different computer vision network models tested on the mosquito picture test set. The test results show that the transformer network model had a higher mosquito species discrimination ability than the CNN.Figure 4Comparison of mosquito recognition effects of computer vision network models and variants. (a) Comparison of mosquito identification accuracy between 3 CNNs and 3 Transformer; (b) The best effect CNN (YOLOv5) training set loss curve(blue), validation set loss curve(green) and validation set accuracy curve(orange); (c) The best effect Transformer (Swin Transformer) training set loss curve, validation set loss curve and validation set accuracy curve. (d) Swin-MSI-T test result confusion matrix; (e) Swin-MSI -B test result confusion matrix; (f) Swin-MSI -L test result confusion matrix. Confusion matrix of mosquito labels in which odd numbers represent females and even numbers represent males. The small squares in the confusion matrix represent the recognition readiness rate, from red to green, the recognition readiness rate is getting higher and higher An. sinensis: 1, 2; Cx. pipiens quinquefasciatus: 3, 4; Cx. pipiens pallens: 5, 6; Cx. pipiens molestus: 7,8 Cx. modestus: 9,10; Ae. albopictus: 11, 12 Ae. aegypti: 13, 14; Cx. pallidothorax: 15, 16 Ae. galloisi: 17,18 Ae. vexans: 19, 20; Ae. koreicus: 21, 22 Armigeres subalbatus: 23, 24; Coquillettidia ochracea: 25, 26 Cx. gelidus: 27, 28 Cx. triraeniorhynchus: 29, 30 Mansonia uniformis: 31, 32 An. vagus: 33, 34 Ae. elsaie: 35,36 Toxorhynchites splendens: 37, 38.Full size imageIn the CNN training process (applied to YOLOv5), the validation accuracy requires more than 110 epochs to grow to 0.9, and the validation loss requires 110 epochs to drop to a flat interval; in contrast, during the training step, these losses represent a continuously decreasing process. These results indicate that the deep learning model derived based on the Swin Transformer algorithm was able to achieve a higher recognition accuracy in less time than the rapid convergence ability of the CNN during the iterative process (Fig. 4b).The Swin Transformer model exhibited the highest test accuracy of 96.3%. During the training process, the loss of this model could stabilize after 30 epochs, and its validation accuracy could grow to 0.9 after 20 epochs; during the validation step, the loss can drop to 0.36 after 20 epochs, after which the loss curve fluctuated but did not produce adverse effects (Fig. 4c). Based on the excellent performance of the Swin Transformer model, this model was used as the baseline to carry out the subsequent analyses.Swin Transformer model variant adjustment (2nd round of learning)Following testing performed to clarify the superior performance of the Swin Transformer algorithm, we chose different Drop_path_rate, Embed_dim and Depths parameter settings and labeled the parameter sets as the Swin Transformer-T, Swin Transformer-B, and Swin Transformer-L variants. Drop_path is an efficient regularization method, and an asymmetric Drop_path_rate is beneficial for supervised representation learning when using image classification tasks and Transformer architectures. The Embed_dim parameter represents the image dimensions obtained after the input red–green–blue (RGB) image is calculated by the Swin Transformer block in stage 1. The Depths parameter is the number of Swin Transformer blocks used in the four stages. The parameter information and test results are shown in Table 1. Due to the increase in the Swin Transformer block and Embed_dim parameters in stage 3, the recognition accuracies of the three variants were found to be 95.8%, 96.3%, and 98.2%, Correspondingly, the f1 score were 96.2%, 96.7% and 98.3%; thus, these variants could effectively improve the mosquito species identification ability in a manner similar to the CNN by increasing the number of convolutional channels to extract more features and improve the overall classification ability. In this study, the Swin Transformer-L variant, which exhibited the highest accuracy, was selected as the baseline for the next work.Table 1 Parameters and test accuracy of three variants of Swin Transformer.Full size tableBy plotting a confusion matrix of the test set results derived using the three Swin Transformer variants, we clearly obtained the proportion of correct and incorrect identifications in each category to visually reflect the mosquito species discrimination ability (Fig. 4d–f). In the matrix, the darker diagonal colors indicate higher identification rate accuracies of the corresponding mosquito categories. Among them, five mosquito species were missing because the Ae. vexans, Coquillettidia ochracea, Mansonia uniformis, An. vagus and Toxorhynchites splendens species were represented in the dataset by only females or only males. The confusion matrix shown in Panel C lists the lowest number of mosquito species identification error points and the lowest accuracy level obtained in each category, suggesting that the Swin Transformer-L model has a better classification performance than the Swin Transformer-T and Swin Transformer-B models.Effect of the input image size on the discrimination ability (3rd round of learning)To investigate the relationship between the input image size and mosquito species identification performance, in this study, we conducted a comparison test between input images with sizes of 224 × 224 and 384 × 384, based on the Swin Transformer-L model, and identified 8 categories of mosquito identification accuracy differences. These test results are shown in Table 2. When using an image size of 224 × 224 pixels, the batch_size parameter was set to 16, and when using an image size of 384 × 384 pixels, the batch_size parameter was set to 4; under these conditions, the proportion of utilized video memory accounted for 67%, as shown in Eq. 4, and this was consistent with the description of the relationship between the size of self-attentive operations during the operation of the Swin Transformer model when 384 × 384 pixels images were used. The time required for the Transformer-L model to complete all the training sessions was excessive, reaching 126 h and even exceeding the 124 h required by the YOLOv5 model, which was found to require the highest computation time during the training process in this work. Long-term training process could more fully reflect the performance differences between models. Fortunately and actually, the response speed of the model will not be affected by the training time. Compared to the accuracy of 98.2% obtained for 224 × 224 inputs, the 384 × 384 input image size derived based on the Swin Transformer-L model provided a higher mosquito species identification accuracy of 99.04%, representing an improvement of 0.84%.$$Omega ({text{W}} – {text{MSA}}) = 4{text{HWC}}^{2} + 2{text{M}}^{2} {text{HWC}}$$
(1)
Table 2 Comparison of recognition accuracy for different input image sizes.Full size tableVisualizing and understanding the Swin MSI modelsTo investigate the differences in the attentional features utilized by the Swin MSI and taxonomists for mosquito species identification, we applied the Grad-CAM method to visualize the Swin MSI attentional areas on mosquitoes at different stages. Because the Swin Transformer has different attentional ranges among its multi-head self-attention steps in different stages, different attentional weights can be found on different mosquito positions. In stage 1, the feature dimension of each patch was 4 × 4 × C, thus enabling the Swin Transformer’s multi-head self-attention mechanism to give more attention to the detailed parts of the mosquitoes, such as their legs, wings, antennae, and pronota. In stage 2, the feature dimension of each patch was 8 × 8 × 2C, enabling the Swin Transformer’s multi-head self-attention mechanism to focus on the bodies of the mosquitoes, such as their heads, thoraces, and abdomens. In stage 3, when the feature dimension of each patch was 16 × 16 × 4C, the Swin Transformer’s multi-head self-attention mechanism could focus on most regions of the mosquito, thus forming a global overall attention mechanism for each mosquito (Fig. 5). This attentional focus process is essentially the same as the process used by taxonomists when classifying mosquito morphology, changing from details to localities to the whole mosquito.Figure 5Attention visualization of representative mosquitoes of the genera Ae., Cx., An., Armigeres, Coquillettidia and Mansonia. This is a visualization for identifying the regions in the image that can explain the classification progress. Images of Toxorhynchites contain only males, with obvious differences in morphological characteristics, are not shown.Full size imageAe. aegypti is widely distributed in tropical and subtropical regions around the world and transmits Zika, dengue and yellow fever. A pair of long-stalked sickle-shaped white spots on both shoulder sides of the mesoscutum, with a pair of longitudinal stripes running through the whole mesotergum, is the most important morphological identification feature of this species. This feature was the deepest section in the attention visualization, indicating that the Swin MSI model also recognized it as the principal distinguishing feature. In addition, the abdominal tergum of A. aegypti is black and segments II-VII have lateral silvery white spots and basal white bands; the model also focused on these areas.Cx. triraeniorhynchus is the main vector of Japanese encephalitis; this mosquito has a small body size, a distinctive white ring on the proboscis (its most distinctive morphological feature), and a peppery color on its whole body. Similarly, the model constructed herein focused on both the head and abdominal regions of this species.An. sinensis is the top vector of malaria in China and has no more than three white spots on its anterior wing margin and a distinct white spot on its marginal V5.2 fringe; this feature was observed in Stage 2, at which time the modelstrongly focused on the corresponding area.The most obvious feature of Armigeres subalbatus is the lateral flattening and slightly downward curving of its proboscis; the observation of the attention visualization revealed that the constructed model focused on these regions from Stage 1 to Stage 3. The mesoscutum and abdominal tergum were not critical and were less important for identification than the proboscis, and the attention visualization results correspondingly show that the neural network focused less on these features.Coquillettidia ochracea belongs to the Coquillettidia genus and is golden yellow all over its body, with the most pronounced abdomen among the analyzed species. The model showed a consistent morphological taxonomic focus on the abdomen of this species.Mansonia uniformis is a vector of Malayan filariasis. The abdominal tergum of this species is dark brown, and its abdominal segments II-VII have yellow terminal bands and lateral white spots, which are more obvious than the dark brown feature on proboscis. Through the attention visualization, we determined that the Swin MSI model was more concerned with the abdominal region features than with the proboscis features.Subspecies-level identification tests of mosquitos in the Culex pipiens ComplexFine-grained image classification has been the focus of extensive research in the field of computer vision25,26. Based on the test set (containing 270 images) constructed herein for three subspecies of the Cx. pipiens Complex, the subspecies and sex identification accuracies were 100% when the Swin MSI model was used.The morphological characteristics of Cx. pipiens quinquefasciatus, Cx. pipiens pallens, and Cx. pipiens molestus within the Cx. pipiens Complex are almost indistinguishable, but their host preferences, self-fertility properties, breeding environments, and stagnation overwintering strategies are very different27. Among the existing features available for morphological classification, the stripes on the abdominal tergum of Cx. pipiens quinquefasciatus are usually inverted triangles and are not connected with the pleurosternums, while those of Cx. pipiens pallens are rectangular and are connected with the pleurosternums. Cx. pipiens molestus is morphologically more similar to Cx. pipiens pallens as an ecological subspecies of the Cx. pipiens Complex. However, taxonomists do not recommend using the unstable feature mentioned above as the main taxonomic feature for differentiation. By analyzing the attention visualization results of these three subspecies (last three rows on Fig. 5), we found that the neural networks of Cx. pipiens quinquefasciatus, Cx. pipiens pallens, and Cx. pipiens molestus still focused on the abdominal regions, as shown in dark red. The area of focus of these neural networks differ from that of the human eye, and the results of this study suggest that the Swin MSI model can detect finely granular features among these three mosquito subspecies that are indistinguishable to the naked human eye.Novel mosquito classification attributionAfter we performed a confidence check on the successfully identified mosquito images in the dataset, the lowest confidence value was found to be 85%. A higher confidence threshold mean stricter evaluation criteria, which can better reflect the powerful performance of the model. Therefore, 0.85 was set as the confidence threshold when judging novel mosquitoes. When identifying 10 unknown mosquito species, the highest derived species confidence level was below 85%; when the results were output to the genus level (Fig. 1d), the average probability of obtaining a correct judgment was 96.26%accuracy and 98.09% F1-score (Table 3). The images tested as novel Ae., Cx. and An. mosquito were from Minakshi and Couret et al.28,29.Table 3 Probability of correct attribution of novel species.Full size table More

Seed sampling and germinationSeed from ten trees from each of 21 native Scottish Scots pine populations (Table 1) were collected in March 2007 and germinated at the James Hutton Institute, Aberdeen (latitude 57.133214, longitude −2.158764) in June 2007. Populations were chosen to represent the species’ native range in Scotland and to include three populations from each of the seven seed zones (Fig. 2). There was no selection of seed-trees on the basis of any traits except for the possession of cones on the date of sampling. Ten seed trees were sampled from each population according to a spatial protocol designed to cover a circle of approximately 1 km in diameter located around a central tree. The sampling strategy identified nine points each in a pre-determined random direction from the central point, whilst stratifying the number sampled with increasing distance from the central point in the ratio 1: 3: 5. This strategy avoids over-sampling the areas close to the centre point. For smaller fragments of woodland, or where only a few trees with cones were present, then the directions of the sampled trees from the central tree were maintained to give a wide coverage of the woodland area, but the distances between trees varied but were never closer than 50 m. To break dormancy, seeds were soaked for 24 hours on the benchtop at room temperature, after which they were stored in wet paper towels and refrigerated in darkness at 3–5 °C for approximately 4 weeks. Seeds were kept moist and transferred to room temperature until germination began (approx. 5–7 days), then transplanted to 8 cm × 8 cm × 9 cm, 0.4 L pots filled with Levington’s C2a compost and 1.5 g of Osmocote Exact 16–18 months slow release fertiliser and kept in an unheated glasshouse. The compost was covered with a layer of grit to reduce moss and liverwort growth. Seedlings from the same mother tree are described as a family and are assumed to be half-siblings.Table 1 Locations and basic environmental data for the populations sampled for seed to establish the trial. See the maternal traits dataset15 for precise data for each mother tree sampled.Full size tableExperimental design: nurseriesThe full collection consisted of 210 families (10 families from each of 21 populations) each consisting of 24 half sibling progeny (total 5,040 individuals); needle tissue was sampled from each seedling and preserved for long term storage, one needle on silica gel, 2–5 needles at −20 °C. After transfer into pots, 8 seedlings per family were moved to one of three nurseries (total 1,680 seedlings per nursery): outdoors at Inverewe Gardens in western Scotland (nursery in the west of Scotland: coded NW, latitude 57.775714, longitude −5.597181, Fig. 2); outdoors in a fruit cage (to minimise browsing) at the James Hutton Institute in Aberdeen (nursery in the east of Scotland: NE); in an unheated glasshouse at the James Hutton Institute in Aberdeen (nursery in a glasshouse: NG). Trees were arranged in 40 randomised trays (blocks) in each nursery. Each block contained two trees per population (total 42 trees). Watering was automatic in NG, and manually as required for NE and NW. No artificial light was used in any of the nurseries. In May 2010 the seedlings from NG were moved outdoors to Glensaugh in Aberdeenshire (latitude 56.893567, longitude −2.535736). In 2010 all plants were repotted into 19 cm (3 L) pots containing Levingtons CNSE Ericaceous compost with added Osmocote STD 16–18 month slow release fertilizer.Experimental design: field sitesIn 2012 the trees were transplanted to one of three field sites: Yair in the Scottish Borders (field site in the south of Scotland: FS, latitude 55.603625, longitude −2.893025); Glensaugh (field site in the east of Scotland: FE); and Inverewe (field site in the west of Scotland: FW). All trees transplanted to FS were raised in the NG and all but four of the trees transplanted to FE were raised locally in the NE (the remainder were grown in NG). In contrast, following mortality and ‘beating up’ (filling gaps where saplings had died), the FW experiment ultimately contained cohorts of trees raised in each of the three nurseries as follows: 290 grown locally in the NW; 132 were grown in the NG; and 82 were grown in the NE.Site historiesThe Yair site (FS) had previously been used for growing Noble fir (Abies procera) for Christmas trees and Lodgepole pine (Pinus contorta), a section of the former were felled and chipped to create a clear area prior to planting. The planting site is also adjacent to a large block of commercial Sitka spruce (Picea sitchensis) forestry, and the Glenkinnon Burn Site of Special Scientific Interest (SSSI NatureScot site code 736; EU site code 135445), an area of mixed broadleaf woodland. Prior to planting, major areas of tall weeds were strimmed. The site was protected by a deer fence. The experiment was planted 8–11 October 2012. The Glensaugh site (FE) is in Forestry Compartment 3 of the Glensaugh Research Station, adjacent to Cleek Loch. It is thought to have been cleared of Scots pine and Larch (Larix decidua) around 1917, after which it reverted to rough grazing. An attempt to reseed part of the site in the 1980s was unsuccessful and it quickly reverted to rough grazing for a second time. The whole site (within which the experimental area is embedded) was deer fenced and re-planted under the Scottish Rural Development Programme (SRDP) in 2012. The experimental plot was planted up 7–9 March 2012. The Inverewe site (FW) had previously been a Sitka spruce and Lodgepole pine plantation (50:50 mix) that had been clear-felled in 2010 following substantial windthrow. The experimental site was deer fenced in early 2012, and the experiment was planted 12–16 March 2012, followed by beating up on 27–28 March 2013 and 22–24 October 2013. There had been minimal preparation of the site in line with current practice for restocking sites. The experimental site is included in the Inverewe Forest Plan, which included deer fencing of a larger area (2014) around the experimental site. Planting of this area was completed in early 2015, funded by NTS (National Trust for Scotland), although natural regeneration is also taking place.At each site, trees were planted in randomised blocks at 3 m × 3 m spacing. There are four randomised blocks in both FS and FE and three in FW. A guard row of Scots pine trees was planted around the periphery of the blocks and between blocks B and C at FS. Each block comprised one individual from each of eight (of the 10 sampled) families per 21 populations (168 trees). Although most families (N = 159) were represented at each of the three sites, families with insufficient trees (N = 9) were replaced in one site (FS) with a different family from the same population. Each experimental site was designed with redundancy such that, if thinning becomes necessary as the trees mature, then the systematic removal of trees (i.e. trees 1,3,5,7, etc of row 1, and 2,4,6,8, etc of row 2, 1,3,5,7,etc of row 3) will maintain a balanced design of the experiment, with sufficient family and population representation to provide an ongoing experiment with full geographic coverage.The field sites generally experience different climates, with FW typically warmer and wetter and with more growing degree days per year and a much longer growing season than both FE and FS (Table 2). The coldest site with the shortest growing season is generally FE.Table 2 Average climatic variables at field sites Glensaugh (FE), Inverewe (FW) and Yair (FS) from planting in 2012 until 2020. Climatic variables are derived from data provided by the Met Office (daily mean, minimum and maximum temperatures and monthly rainfall).Full size tablePhenotype assessmentsMaternal traitsFollowing seed collection, a range of traits were measured in the mother trees in order to control for maternal effects in subsequent measurements of their progeny (Table 3). For each mother tree, measurements of height and diameter at breast height (DBH) were taken, and ten cones were collected and assessed in detail. Cone width and length were measured prior to drying the cones (when they were still closed). Cone weight was measured post-drying. Seed removed from each cone was assessed for total weight (after wings had been removed) and for the count and percentage of seeds which were classed as viable (viable seed were those which had both a wing and an obvious seed). No further seed sorting was applied.Table 3 Traits assessed in mother trees, cones, seeds (dataset: Maternal), nursery seedlings (dataset: Nursery) and field trials (dataset: Field). Within the datasets, traits are recorded in a single column for each year using the format Code-Year (e.g. absolute height in 2008 = HA08) except for the maternal traits datasets which were all measured in 2007.Full size tableNursery traitsSeedling phenotype assessments were performed annually from 2007–2010 for three different trait types: phenology (budburst and growth cessation); form (total number of buds, needle length); cumulative growth (stem diameter and height, canopy width). Measurements of tree form and cumulative growth traits were taken after the end of each growing season. Phenology was assessed weekly during the spring and autumn of 2008 for budburst and growth cessation, respectively. Budburst was defined as the number of days from 31 March 2008 to the time when newly emerged green needles were observed (budburst stage 6: Fig. 3). In some seedlings in 2008, a secondary flush of growth occurred from terminal buds that had formed during the summer of that year. Therefore, growth cessation was defined retrospectively as the number of days from 10 September 2008 to the date when a terminal bud had formed on the leading shoot of the seedling, providing no further growth was observed either on the stem below that bud, nor from the bud itself. Canopy width (widest point) was measured at two perpendicular points in the horizontal plane. Needle length was measured for three needles per tree. Mortality was recorded each year from 2007 to 2010.Fig. 3Phenological stages of bud burst in Pinus sylvestris assessed in field trials. Inset numbers correspond to budburst stage. Budburst stage 1: bud dormant; 2: bud swelling and showing signs of linear expansion; 3: scales open at base revealing green tissue. Remaining bud remains encased by smooth bud scales; 4: scales open along length of shoot revealing green tissue and partially visible needles; 5: white tipped needles visible along length of the shoot; 6: green needles emerging away from the shoot (bottle brush appearance) along its entire length; 7: Needles have separated and next year’s terminal bud is usually formed and clearly visible.Full size imageField traitsTree height was measured in the field in the winter after each growing season from 2013 at FE and FW, and from 2014 to 2020 at all sites. Height was taken as the vertical measurement in cm from top bud straight to the ground. Basal stem diameter was measured at the end of the growing season for trees growing at FE and FW from 2014 to 2020 and for FS in 2020.Phenology assessments were performed in spring at each site from 2015 to 2019. Seven distinct stages of budburst (assessed on the terminal bud) were defined (Fig. 3) although only stages 4 to 6 are included in the dataset and considered for analysis due to high proportions of missing data for the early and late stages. Each tree was assessed for budburst stage at weekly intervals from early spring until budburst was complete. In order to allow comparisons within and among sites and years, the date at which each stage of budburst occurred was considered relative to 31 March of that year. For example, 25 May 2019 is recorded as 55 days since 31 March 2019. The duration of budburst (time taken to reach stage 6 from stage 4) was also estimated.When trees progressed through budburst stages rapidly, skipping a stage between assessments, a mean value was taken from the two assessment dates. For example, if a tree was at stage 4 on day 55 and was recorded as stage 6 at the next assessment on day 62, it is assumed to have reached stage 5 at day 58.5. More

Each of the 23 key variables can be used for analysis. To validate the dataset, we used five plant-related variables (diversity of order, family, genus, species and species endemic to China) to demonstrate the process of using the dataset for analysis as follows:(1) For the four variables of plant taxa “order”, “family”, “genus” and “species”, the similarity and difference in spatial distribution pattern of diversity of different taxa in the Qinling-Daba Mountains climate transition zone were analyzed. The spatial distribution pattern of the diversity of the four taxa is shown in Fig. 3, which is increasingly lower from south (low latitude) to north (high latitude). This result is consistent with the classical latitudinal gradient model of plant diversity. The boundary between higher diversity in the south and lower diversity in the north is roughly located in the area of Funiu Mountains in the eastern Qinling-Daba Mountains, Taibai Mountains in the central Qinling-Daba Mountains and Baishui River in the western Qinling-Daba Mountains. However, with the reduction in taxon scale, the spatial distribution pattern of diversity tends to be complex. Orders (Fig. 3a) and families (Fig. 3b) can be divided by lines, while genera (Fig. 3c) need thicker lines, and species (Fig. 3d) can only be divided by polygons. Figure 3 shows that the taxonomic groups of families are more clearly divided, while species can only be divided by staggered bands. Therefore, when dividing the north–south boundary, the family taxon scale is appropriate, whereas the species scale is more appropriate when studying the north–south transition zone.Fig. 3Spatial distribution of diversity of orders, families, genera and species. (a) The blue dotted line is basically the dividing line of the order diversity of 50 species. The order diversity to the north of the blue dotted line is lower than 50 species, and the order diversity to the south of the blue dotted line is higher than 50 species. (b) The blue dotted line is basically the dividing line of the family diversity of 150 species. The family diversity to the north of the blue dotted line is lower than 150 species, and the family diversity to the south of the blue dotted line is higher than 150 species. (c) The thicker blue dotted line is basically the dividing line of genus diversity of 578–681 species. The genus diversity to the north of the blue dotted line is lower than 578 species, and the genus diversity to the south of the blue dotted line is higher than 681 species. (d) The blue area is basically the dividing line of species diversity of 1385–1618 species. The species diversity to the north of the blue dotted line is lower than 1385 species, and the species diversity to the south of the blue dotted line is higher than 1618 species.Full size imageThe dataset can also count the orders, families and genera that appear in 58 nature reserves, indicating that these orders, families and genera are widely distributed in this area, while the orders, families and genera that only appear in a single nature reserve indicate that these taxa are unique to this nature reserve in this area, reflecting their locality and uniqueness, which is helpful to understanding the specific distribution of plants in detail. The relevant statistics are as follows:
There are 28 orders present in every nature reserve:
Liliales, Dipsacales, Lamiales, Fabales, Ericales, Poales, Saxifragales, Malpighiales, Malvales, Asterales, Fagales, Gentianales, Geraniales, Ranunculales, Rosales, Solanales, Apiales, Cornales, Brassicales, Caryophyllales, Dioscoreales, Santalales, Myrtales, Asparagales, Celastrales, Sapindales, Alismatales, and Boraginales.The order that only appears in one nature reserve is Petrosaviales, which appears in the Dabashan Nature Reserve in Chongqing.
There are 51 families present in every nature reserve:
Liliaceae, Primulaceae, Plantaginaceae, Lamiaceae, Euphorbiaceae, Cannabaceae, Juncaceae, Fabaceae, Poaceae, Elaeagnaceae, Betulaceae, Apocynaceae, Violaceae, Malvaceae, Crassulaceae, Campanulaceae, Asteraceae, Orchidaceae, Polygonaceae, Orobanchaceae, Onagraceae, Gentianaceae, Geraniaceae, Ranunculaceae, Rubiaceae, Rosaceae, Caprifoliaceae, Thymelaeaceae, Apiaceae, Cyperaceae, Cornaceae, Paeoniaceae, Brassicaceae, Amaryllidaceae, Caryophyllaceae, Rhamnaceae, Santalaceae, Asparagaceae, Celastraceae, Sapindaceae, Adoxaceae, Araliaceae, Berberidaceae, Hydrangeaceae, Scrophulariaceae, Convolvulaceae, Urticaceae, Salicaceae, Papaveraceae, Iridaceae, and Boraginaceae.There are 15 families that only appear in one nature reserve, as shown in Table 2.Table 2 Endemic families of the nature reserves in the Qinling-Daba Mountains and surrounding areas.Full size table
There are 54 genera present in every nature reserve:
Patrinia, Polygonum, Sanicula, Plantago, Allium, Delphinium, Euphorbia, Juncus, Cynanchum, Trigonotis, Artemisia, Sorbus, Polygonatum, Scutellaria, Cirsium, Viburnum, Ajuga, Viola, Galium, Geranium, Salix, Epilobium, Gentiana, Ranunculus, Malus, Acer, Rubia, Rosa, Torilis, Lonicera, Adenophora, Philadelphus, Cornus, Paeonia, Rhamnus, Rumex, Carex, Thalictrum, Asparagus, Carpesium, Clematis, Potentilla, Euonymus, Eleutherococcus, Berberis, Spiraea, Rubus, Populus, Vicia, Silene, Iris, Poa, Aster, and Buddleja.There were 225 genera that only appeared in one nature reserve, as shown in Figshare file 269.(2) For the “species endemic to China” variable of plants, we can see from the diversity distribution pattern of species endemic to China in this region (Fig. 4) that the number of endemic species in the Qinling-Daba Mountains is higher than that of species outside of the region, which reflects the strong transition zone in the Qinling-Daba Mountains. The variables of species endemic to China obtained from the Qinling-Daba Mountains and their surroundings were clustered by the Bray–Curtis dissimilarity measure70 and Ward’s minimum variance (the clustering method recommended for plant cluster analysis). The clustering results are shown in Fig. 5a. At the same time, the clustering results are displayed in space. Figure 5b shows that category 3 extends from the east outside the Qinling-Daba Mountains to the Baishuijiang Nature Reserve inside the western Qinling-Daba Mountains, which is consistent with the fact that the Qinling-Daba Mountains are an important ecogeographical “corridor” connecting the east and the west.Fig. 4Spatial distribution of diversity of species endemic to China in the Qinling-Daba Mountains and adjacent areas.Full size imageFig. 5(a) Clustering results of Ward’s connection aggregation of species endemic to China in 58 nature reserves. (b) Spatial distribution of clustering results of species endemic to China; the larger the dot and the darker the color, the earlier it is merged into this category, and the smaller the dot and the lighter the color, the later it is merged into this category.Full size image More

Mueller, R. C. et al. An emerging view of the diversity, ecology, and function of Archaea in alkaline hydrothermal environments. FEMS Microbiol. Ecol. 97, fiaa246 (2020).
Google Scholar 
López-López, O., Cerdán, M.-E. & González-Siso, M.-I. Thermus thermophilus as a source of thermostable lipolytic enzymes. Microorganisms 3, 792–808 (2015).PubMed 
PubMed Central 

Google Scholar 
Sahay, H. et al. Hot springs of Indian Himalayas: Potential sources of microbial diversity and thermostable hydrolytic enzymes. 3 Biotech 7, 118 (2017).PubMed 
PubMed Central 

Google Scholar 
Patel, A. K., Singhania, R. R., Sim, S. J. & Pandey, A. Thermostable cellulases: Current status and perspectives. Bioresour Technol 279, 385–392 (2019).CAS 
PubMed 

Google Scholar 
Decastro, M.-E., Rodríguez-Belmonte, E. & González-Siso, M.-I. Metagenomics of thermophiles with a focus on discovery of novel thermozymes. Front. Microbiol. 7, 1521–1521 (2016).PubMed 
PubMed Central 

Google Scholar 
Meslé, M. M. et al. Isolation and characterization of lignocellulose-degrading geobacillus thermoleovorans from Yellowstone National Park. Appl. Environ. Microbiol. 88, e0095821 (2022).PubMed 

Google Scholar 
Verma, P., Yadav, A. N., Shukla, L., Saxena, A. K. & Suman, A. Hydrolytic enzymes production by thermotolerant Bacillus altitudinis IARI-MB-9 and Gulbenkiania mobilis IARI-MB-18 isolated from Manikaran hot springs. Int. J. Adv. Res. 3, 1241–1250 (2015).CAS 

Google Scholar 
Wu, B. et al. Microbial sulfur metabolism and environmental implications. Sci. Total Environ. 778, 146085 (2021).ADS 
CAS 
PubMed 

Google Scholar 
Lavrentyeva, E. V. et al. Bacterial diversity and functional activity of microbial communities in hot springs of the Baikal Rift Zone. Microbiology 87, 272–281 (2018).CAS 

Google Scholar 
Miller Scott, R., Strong Aaron, L., Jones Kenneth, L. & Ungerer Mark, C. Bar-Coded pyrosequencing reveals shared bacterial community properties along the temperature gradients of two alkaline hot springs in Yellowstone National Park. Appl. Environ. Microbiol. 75, 4565–4572 (2009).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Sharp, C. E. et al. Humboldt’s spa: Microbial diversity is controlled by temperature in geothermal environments. ISME J. 8, 1166–1174 (2014).CAS 
PubMed 
PubMed Central 

Google Scholar 
Stefanova, K. et al. Archaeal and bacterial diversity in two hot springs from geothermal regions in Bulgaria as demostrated by 16S rRNA and GH-57 genes. Int. Microbiol. 18, 217–223 (2015).CAS 
PubMed 

Google Scholar 
Hou, W. et al. A comprehensive census of microbial diversity in hot springs of Tengchong, Yunnan Province China using 16S rRNA gene pyrosequencing. PLoS ONE 8, e53350 (2013).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Sahm, K. et al. High abundance of heterotrophic prokaryotes in hydrothermal springs of the Azores as revealed by a network of 16S rRNA gene-based methods. Extremophiles 17, 649–662 (2013).CAS 
PubMed 

Google Scholar 
Purcell, D. et al. The effects of temperature, pH and sulphide on the community structure of hyperthermophilic streamers in hot springs of northern Thailand. FEMS Microbiol. Ecol. 60, 456–466 (2007).CAS 
PubMed 

Google Scholar 
Meyer-Dombard, D. R. & Amend, J. P. Geochemistry and microbial ecology in alkaline hot springs of Ambitle Island, Papua New Guinea. Extremophiles 18, 763–778 (2014).CAS 
PubMed 

Google Scholar 
de Leon, K. B., Gerlach, R., Peyton, B. M. & Fields, M. W. Archaeal and bacterial communities in three alkaline hot springs in Heart Lake Geyser Basin, Yellowstone National Park. Front. Microbiol. 4, 10 (2013).
Google Scholar 
Boomer, S. M., Noll, K. L., Geesey, G. G. & Dutton, B. E. Formation of multilayered photosynthetic biofilms in an alkaline thermal spring in Yellowstone National Park, Wyoming. Appl. Environ. Microbiol. 75, 2464–2475 (2009).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Wang, S. et al. Greater temporal changes of sediment microbial community than its waterborne counterpart in Tengchong hot springs, Yunnan Province, China. Sci. Rep. 4, 7479 (2014).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Sun, Y., Liu, Y., Pan, J., Wang, F. & Li, M. Perspectives on cultivation strategies of archaea. Microb. Ecol. 79, 770–784 (2020).PubMed 

Google Scholar 
Brock, T. D. Life at high temperatures. Science 158, 1012 (1967).ADS 
CAS 
PubMed 

Google Scholar 
Christiansen, R. L. The Quaternary and Pliocene Yellowstone Plateau volcanic field of Wyoming, Idaho, and Montana. Professional Paper (2001).Rowe, J. J., Fournier, R. & Morey, G. Chemical analysis of thermal waters in Yellowstone National Park, Wyoming, 1960–65. USGS https://doi.org/10.3133/b1303 (1973).Article 

Google Scholar 
Fournier, R., Thompson, M. J. & Hutchinson, R. A. The geochemistry of hot spring waters at Norris Geyser Basin, Yellowstone National Park. International symposium on water-rock interactions (1992).Podar, P. T., Yang, Z., Björnsdóttir, S. H. & Podar, M. Comparative analysis of microbial diversity across temperature gradients in hot springs from Yellowstone and Iceland. Front. Microbiol. 11, 1625 (2020).PubMed 
PubMed Central 

Google Scholar 
Pala, C. et al. Environmental drivers controlling bacterial and archaeal abundance in the sediments of a Mediterranean lagoon ecosystem. Curr. Microbiol. 75, 1147–1155 (2018).CAS 
PubMed 
PubMed Central 

Google Scholar 
Foyer, C. H., Noctor, G. & Hodges, M. Respiration and nitrogen assimilation: Targeting mitochondria-associated metabolism as a means to enhance nitrogen use efficiency. J. Exp. Bot. 62, 1467–1482 (2011).CAS 
PubMed 

Google Scholar 
Ershanovich, V. N. et al. Nitrogen assimilation enzymes in Bacillus subtilis mutants with hyperproduction of riboflavin. Mol. Gen. Mikrobiol. Virusol. 2005(3), 29–34 (2005).
Google Scholar 
Offre, P., Spang, A. & Schleper, C. Archaea in biogeochemical cycles. Annu Rev Microbiol 67, 437–457 (2013).CAS 
PubMed 

Google Scholar 
Cabello, P., Roldán, M. D. & Moreno-Vivián, C. Nitrate reduction and the nitrogen cycle in archaea. Microbiology 150, 3527–3546 (2004).CAS 
PubMed 

Google Scholar 
Graupner, M., Xu, H. & White, R. H. The pyrimidine nucleotide reductase step in riboflavin and F(420) biosynthesis in archaea proceeds by the eukaryotic route to riboflavin. J. Bacteriol. 184, 1952–1957 (2002).CAS 
PubMed 
PubMed Central 

Google Scholar 
Chernyh, N. A. et al. Dissimilatory sulfate reduction in the archaeon “Candidatus Vulcanisaeta moutnovskia” sheds light on the evolution of sulfur metabolism. Nat. Microbiol. 5, 1428–1438 (2020).CAS 
PubMed 

Google Scholar 
Castelle, C. J. & Banfield, J. F. Major new microbial groups expand diversity and alter our understanding of the tree of life. Cell 172, 1181–1197 (2018).CAS 
PubMed 

Google Scholar 
Williams, T. A. et al. Integrative modeling of gene and genome evolution roots the archaeal tree of life. Proc. Natl. Acad. Sci. U.S.A. 114, E4602–E4611 (2017).ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Guy, L. & Ettema, T. J. G. The archaeal ‘TACK’ superphylum and the origin of eukaryotes. Trends Microbiol. 19, 580–587 (2011).CAS 
PubMed 

Google Scholar 
Wang, Y., Wegener, G., Hou, J., Wang, F. & Xiao, X. Expanding anaerobic alkane metabolism in the domain of Archaea. Nat. Microbiol. 4, 595–602 (2019).CAS 
PubMed 

Google Scholar 
Hedlund, B. P. et al. Uncultivated thermophiles: Current status and spotlight on ‘Aigarchaeota’. Curr. Opin. Microbiol. 25, 136–145 (2015).CAS 
PubMed 

Google Scholar 
Reichart, N. J. et al. Activity-based cell sorting reveals responses of uncultured archaea and bacteria to substrate amendment. ISME J. 14, 2851–2861 (2020).CAS 
PubMed 
PubMed Central 

Google Scholar 
Hua, Z.-S. et al. Genomic inference of the metabolism and evolution of the archaeal phylum Aigarchaeota. Nat. Commun. 9, 2832 (2018).ADS 
PubMed 
PubMed Central 

Google Scholar 
Beam, J. P. et al. Ecophysiology of an uncultivated lineage of Aigarchaeota from an oxic, hot spring filamentous “streamer” community. ISME J. 10, 210–224 (2016).CAS 
PubMed 

Google Scholar 
Gonsior, M. et al. Yellowstone hot springs are organic chemodiversity hot spots. Sci. Rep. 8, 14155 (2018).ADS 
PubMed 
PubMed Central 

Google Scholar 
Gibson, M. L. & Hinman, N. W. Mixing of hydrothermal water and groundwater near hot springs, Yellowstone National Park (USA): Hydrology and geochemistry. Hydrogeol. J. 21, 919–933 (2013).ADS 
CAS 

Google Scholar 
Campbell, K. M. et al. Sulfolobus islandicus meta-populations in Yellowstone National Park hot springs. Environ. Microbiol. 19, 2334–2347 (2017).PubMed 

Google Scholar 
Thiel, V. et al. The dark side of the mushroom spring microbial mat: Life in the shadow of chlorophototrophs. I. Microbial diversity based on 16S rRNA gene amplicons and metagenomic sequencing. Front. Microbiol. 7, 919 (2016).PubMed 
PubMed Central 

Google Scholar 
Caporaso, J. G. et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proc. Natl. Acad. Sci. U.S.A. 108, 4516–4522 (2011).ADS 
CAS 
PubMed 

Google Scholar 
Parada, A. E., Needham, D. M. & Fuhrman, J. A. Every base matters: Assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environ. Microbiol. 18, 1403–1414 (2016).CAS 
PubMed 

Google Scholar 
Apprill, A., McNally, S., Parsons, R. & Weber, L. Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquat. Microb. Ecol. 75, 129–137 (2015).
Google Scholar 
Thompson, L. R. et al. A communal catalogue reveals Earth’s multiscale microbial diversity. Nature 555, 457–463 (2017).ADS 

Google Scholar 
Eloe-Fadrosh, E. A., Ivanova, N. N., Woyke, T. & Kyrpides, N. C. Metagenomics uncovers gaps in amplicon-based detection of microbial diversity. Nat. Microbiol. 1, 15032 (2016).CAS 
PubMed 

Google Scholar 
Edgar, R. C. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461 (2010).CAS 
PubMed 

Google Scholar 
Edgar, R. C. UNOISE2: Improved error-correction for Illumina 16S and ITS amplicon sequencing. BioRxiv https://doi.org/10.1101/081257 (2016).Article 

Google Scholar 
Murali, A., Bhargava, A. & Wright, E. S. IDTAXA: A novel approach for accurate taxonomic classification of microbiome sequences. Microbiome 6, 140 (2018).PubMed 
PubMed Central 

Google Scholar 
Parks, D. H. et al. A complete domain-to-species taxonomy for Bacteria and Archaea. Nat. Biotechnol. 38, 1079–1086 (2020).CAS 
PubMed 

Google Scholar 
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).CAS 
PubMed 
PubMed Central 

Google Scholar 
Stamatakis, A. RAxML version 8: A tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313 (2014).CAS 
PubMed 
PubMed Central 

Google Scholar 
Letunic, I. & Bork, P. Interactive Tree Of Life (iTOL) v5: An online tool for phylogenetic tree display and annotation. Nucleic Acids Res. 49, W293–W296 (2021).CAS 
PubMed 
PubMed Central 

Google Scholar 
Matsen, F. A., Kodner, R. B. & Armbrust, E. V. pplacer: Linear time maximum-likelihood and Bayesian phylogenetic placement of sequences onto a fixed reference tree. BMC Bioinform. 11, 538 (2010).
Google Scholar 
Chong, J., Liu, P., Zhou, G. & Xia, J. Using MicrobiomeAnalyst for comprehensive statistical, functional, and meta-analysis of microbiome data. Nat. Protoc. 15, 799–821 (2020).CAS 
PubMed 

Google Scholar 
Chambers, M. C. et al. A cross-platform toolkit for mass spectrometry and proteomics. Nat. Biotechnol. 30, 918–920 (2012).CAS 
PubMed 
PubMed Central 

Google Scholar 
Pluskal, T., Castillo, S., Villar-Briones, A. & Oresic, M. MZmine 2: Modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular profile data. BMC Bioinform. 11, 395 (2010).
Google Scholar 
Patiny, L. & Borel, A. ChemCalc: A building block for tomorrow’s chemical infrastructure. J. Chem. Inf. Model. 53, 1223–1228 (2013).CAS 
PubMed 

Google Scholar 
Chong, J., Wishart, D. S. & Xia, J. Using MetaboAnalyst 4.0 for comprehensive and integrative metabolomics data analysis. Curr. Protoc. Bioinform. 68, e86 (2019).
Google Scholar 
Liu, G., Lee, D. P., Schmidt, E. & Prasad, G. L. Pathway analysis of global metabolomic profiles identified enrichment of caffeine, energy, and arginine metabolism in smokers but not moist snuff consumers. Bioinform. Biol. Insights 13, 1177932219882961–1177932219882961 (2019).PubMed 
PubMed Central 

Google Scholar 
Xia, J. & Wishart, D. S. MetPA: A web-based metabolomics tool for pathway analysis and visualization. Bioinformatics 26, 2342–2344 (2010).CAS 
PubMed 

Google Scholar 
Huber, W. et al. Orchestrating high-throughput genomic analysis with Bioconductor. Nat. Methods 12, 115–121 (2015).CAS 
PubMed 
PubMed Central 

Google Scholar 
Rohart, F., Gautier, B., Singh, A. & Lé Cao, K.-A. mixOmics: An R package for ’omics feature selection and multiple data integration. PLoS Comput. Biol. 13, e1005752–e1005752 (2017).ADS 
PubMed 
PubMed Central 

Google Scholar  More

The richness of Melanesian FrogsApproximately 7.2% (534 out of 7404) of Earth’s recognised frog species occur in Melanesia, a region comprising < 0.7% of the world’s land area. Frog richness in Melanesia, and especially on New Guinea and nearby land-bridge islands (471 species), is higher than in any other tropical insular region (Fig. 1a). New Melanesian frog species have been described at an average rate of nearly 13 species/year since 2000, and the recognised frog fauna has grown by > 50% in that timeframe (Fig. 1b). The authorship of new species has been concentrated, with six authors featuring on 20 or more descriptions since 2000, and one or more of these six authors on every species description since 2000. A small number of species descriptions has included genetic data (31 species), although a higher number of Melanesian frog species have at least one sequence available on GenBank (~38%, or approximately 200 species). This taxonomic work has revealed or emphasised many evolutionary novelties (Fig. 2): multiple apparently independent derivations of extremely miniaturised vertebrates22,23,24, including some of the world’s smallest known tetrapods23,25,26; multiple derivations of complex parental care in different genera27,28; frequent evolutionary shifts between terrestrial, arboreal and scansorial lifecycles22,29; the most extreme sexual size dimorphism yet documented in anurans30; drastic ontogenetic colour change31; a radiation of canopy-dwelling treefrogs32 that show extensive finger webbing and parachuting behaviour convergent with unrelated frog lineages in Asia and the Neotropics; and treefrogs with erectile noses33,34. Taxonomic work has also elucidated novel concentrations of range-restricted endemic taxa, especially in the Milne Bay Region at the far eastern edge of New Guinea21.Fig. 1: Temporal trends in the documentation of the Melanesian frog fauna.a Species accumulation curves for species-rich ( >100 species) insular frog biotas (Species lists from AmphibiaWeb as of 1 October 2021). Separate accumulation curves are given for the entire fauna of Melanesia (including New Guinea), and the fauna of New Guinea and nearby predominantly land-bridge islands. b Species accumulation curve for frogs within Melanesia. Bar at end indicates predicted number of species in each major family based on known, but as yet undescribed candidate species.Full size imageFig. 2: Melanesian frog species described within the last 15 years illustrating the ecological and morphological diversity of the fauna.a Paedophryne titan and b Choerophryne gracilirostris – examples of lineages that have undergone convergent minaturisation; c Choerophryne alpestris – a fossorial species within a largely scansorial lineage; d Xenorhina macrodisca – scansorial species within a largely fossorial lineage; e Cornufer custos and f Oreophryne oviprotector – independent derivations of complex parental care; g Litoria pallidofemora – extensive digital webbing for parachuting; and h Litoria pinocchio – sexually dimorphic and erectile rostral spikes. Photographs F. Kraus (a), S. Richards (b–g), and courtesy of T. Laman (h).Full size imageFrog species richness in Melanesia is highly concentrated into just three families, with Pelodryadidae (137 recognised species, estimated ~200) and especially Microhylidae (317 species, estimated >400) dominating. Melanesian Pelodryadidae are phylogenetically interdigitated with relatives in Australia, suggesting multiple dispersals between the two regions35. In contrast, ancestors of the direct-developing microhylids colonised Melanesia from Asia via trans-marine dispersal likely only once36, radiated across open ecological niches37, and are now the most species-rich insular radiation of frogs in the world. The third major family comprises an ecologically diverse radiation of the direct-developing Ceratobatrachidae (57 species, estimated 66) largely associated with island-arc terranes of East Melanesia and the Philippines, indicating a long history of insular diversification and trans-marine dispersal38. The predominance of direct-developing frogs in Melanesia (~70% of species) mirrors insular faunas in Madagascar (~34%), Sri Lanka (~67%) and the Greater Antilles (~87%). The other four frog families in Melanesia are all relatively species poor (2, 3, 4, and 13 species) (Fig. 1a), centred in New Guinea, and include lineages originating in Asia (Ranidae, Dicroglossidae) or Australia (Myobatrachidae, Limnodynastidae).The described diversity of Melanesian amphibian species remains an underestimate. Survey work and investigation of museum collections by the co-authors identified ~190 additional candidate species distributed across 16 different genera, mostly from Papua New Guinea, suggesting a total richness of over 700 frog species (Fig. 1a, Supplementary Table 1). This estimated percentage of undescribed diversity (~25%) mirrors estimates for the New Guinean flora (~18–22%)7. The majority of candidate species are concentrated in the two most diverse families (Microhylidae and Pelodryadidae), although genetic, morphological, and acoustic evidence indicate the diversity of Melanesian Ranidae is also underestimated (S. Richards and F. Kraus pers. obs.). Most material documenting candidate species has been collected in the last 20 years, and the vast majority is from Papua New Guinea (Supplementary Fig. 1). There is some suggestion of a slowing in the rate of candidate species discovery in the last decade (Supplementary Fig. 2); however, several of the most active field workers in this region have ceased survey work in recent years, which likely accounts for much of this decline. The pervasiveness of complexes of morphologically and/or acoustically cryptic taxa is poorly understood; survey work continues to reveal novelties, and large areas of the region remain unsurveyed or undersampled. In particular, comparisons of area-to-diversity ratios between the better-known eastern portion of New Guinea (Papua New Guinea) with the poorly surveyed western (Indonesian) portion of the island further suggest that, even with candidate species included, diversity in the latter region may be underestimated by as much as 50% (Supplementary Methods and Results, Supplementary Table 2). These trends and patterns all indicate that ~ 700 species is a very conservative minimum estimate of total diversity and support analyses in other taxa showing Melanesia remains a hotspot of unrecognised diversity39,40.Endemism and distributional patternsThe Melanesian frog fauna is highly endemic (97.2%), with tiny proportions of species shared with Australia (2.4%) or with islands farther west in Indonesia (0.6%), indicating that Australia and Melanesia are discrete centres of frog diversification, despite periodic connection via land bridges through the late Tertiary41. The vast majority of Melanesian frog species (471) occur on New Guinea and nearby land-bridge islands (Raja Ampats, Japen and the Milne Bay islands). In comparison, the frog fauna of the much smaller region of Maluku is depauperate (16 species, of which nine are endemic) but also almost certainly underestimated (e.g., there are no Microhylidae recorded from Buru). Most taxa from Maluku are congeneric (and several conspecific) with lineages centred on New Guinea, supporting the biogeographic clustering of Maluku’s amphibians with the main island of New Guinea. In contrast, the frog fauna of East Melanesia is more diverse and highly endemic and dominated by an ecologically diverse radiation of a different family (Ceratobatrachidae) with only four (all pelodryadid treefrogs) out of 56 species shared with nearby New Guinea. East Melanesia and New Guinea appear to be discrete and long-isolated centres of diversification, as expected from their independent geological histories42.Melanesia spans five countries, and this has possibly to some degree masked the exceptional species diversity of the overall region. Papua New Guinea has the highest number of species (398) and endemic species (318). This likely reflects some combination of its slightly larger area (when islands to the north are included), more diverse geological origins, and greater inventory work than seen in neighbouring regions of Indonesia7. Papua, West Papua and Maluku (Indonesia) have many fewer documented species (197), of which a majority (134) is endemic. The boundary between Papua and Papua New Guinea is visible in species-richness maps (Fig. 3a), with lower diversity to the west, indicating that the distribution and diversity of frogs in Indonesia remain less documented in science. The frog faunas of the Solomon Islands (21 species) and Fiji (two species) are more depauperate but include a significant endemic or near-endemic component, whereas the geographically intervening islands of Vanuatu support no native frogs.Fig. 3: Frog species richness in Melanesia based on IUCN distributional maps for all species described by 2019.a All species; b Ceratobatrachidae; c Microhylidae; d Pelodryadidae. Areas of highest estimated diversity correspond to mountain ranges in central and northern New Guinea. The boundaries between Maluku, New Guinea and East Melanesia are indicated. Fiji has only two frog species and is geographically distant from other areas of Melanesia inhabited by frogs and is not visble on this map.Full size imageBased on distribution maps generated for all species recognised by 31 August 2019, the highest regional alpha diversity of frogs occurs along the Central Cordillera of New Guinea (especially in Papua New Guinea) and around the higher mountain ranges along the north coast of Papua New Guinea (Fig. 3a). These centres of diversity correlate with extensive areas of hill and montane forest and broadly correspond with elevational species-richness patterns for mammals and birds in Melanesia43 and for many other taxa elsewhere in the tropics44,45. Large areas of montane forest with lower species richness along the northern versant of the Central Cordillera in Papua New Guinea and in mountain ranges across Papua certainly reflect inadequate sampling. The ceratobatrachid-dominated frog fauna of East Melanesia is richest in Bougainville (Fig. 3d), with attenuating richness towards the west and especially to the east. The two most speciose families both show alpha diversity peaks in mountainous areas of central New Guinea (Fig. 3c–d). In contrast, microhylids are largely absent from the seasonally dry woodlands of the Trans-Fly region in southern New Guinea and exhibit high diversity in northern New Guinea, whereas pelodryadids are much more speciose in the lowlands of southern New Guinea than northern New Guinea. These broad trends may have both ecological (sensitivity of direct-developing microhylids to dry conditions) and historical (Australia as a centre of origin for savanna-adapted Pelodrydidae) underpinnings.The historical and contemporary factors underpinning high frog species diversity in New Guinea remain largely unstudied, especially when compared to other species-rich insular amphibian faunas such as Madagascar46 or the Greater Antilles47. When compared to some areas of the Neotropics, alpha and beta diversities of frogs in lowland forests in the basins of the Sepik and Ramu rivers in New Guinea are unremarkable48. However, the Milne Bay Region has exceptionally high levels of endemism21, so species turnover will be higher in this area. Extent-of-occurrence estimates derived from IUCN maps indicate that direct-developing microhylids have smaller mean and median range sizes than all other families of frogs in Melanesia (Supplementary Table 3). Microhylidae also dominate anuran species diversity in Milne Bay21 and many mountain areas where standing water is very limited49. These data suggest that, as with some areas in the Neotropics50, high beta diversity in lineages with direct development is a key factor underpinning amphibian megadiversity in Melanesia. To address these questions further, synthetic analyses are required to better quantify the extent to which regional megadiversity in Melanesia reflects high community diversity versus species turnover, how elevation and insularity moderates these two parameters, and to what extent emergent patterns may differ from diverse frog communities in other regions such as the Neotropics.The conservation status of Melanesian FrogsThe frog fauna of Melanesia is currently less threatened but more Data Deficient than other comparable insular regions (Fig. 4a). The vast majority of Melanesian frogs are categorised as Least Concern (68%) or Data Deficient (24%). Thirty-one species (6%, or 8% if Data Deficient taxa are excluded) are threatened (Critically Endangered, Endangered, Vulnerable) (Supplementary Table 4), and eight species are considered Near Threatened. No species are assessed as Extinct or Extinct in the Wild. Since the first Global Amphibian Assessment in 2004, the number of Melanesian frog species has grown by 44%, and nearly 60% of the 31 Melanesian frog species now considered threatened were described after 2004 (Fig. 1a). Only one change in status between 2004 and 2019 was considered genuine (Cophixalus sphagnicola), due to the emerging threat of a newly opened mine. All other status changes (for 116 taxa) reflect better information on distribution or changed assessment protocols (Supplementary Table 5). Applying stricter criteria for use of the Data Deficient category in the 2019 IUCN assessment reduced the number of Data Deficient species when compared to 2004 (125 versus 197), but Melanesia still has a higher percentage of Data Deficient taxa than other species-rich tropical insular faunas (Fig. 4a).Fig. 4: The conservation status of Melanesian frogs.a Comparison of number of species in each IUCN threat category across Melanesia, other diverse insular regions, and the nearby continent of Australia. Melanesia has a proportionally low number of threatened taxa but high number of Data Deficient taxa (EX Extinct, CR Critically Endangered, EN Endangered, VU Vulnerable, NT Near Threatened, DD Data Deficient, LC Least Concern, NE Not Evaluated); b Slopes around Mt Simpson, Milne Bay Province, a hotspot of threatened frog diversity due to forest loss through conversion to anthropogenic grasslands; c Choerophryne sanguinopicta from Mt Simpson (Critically Endangered); d Oreophryne ezra from Rossel Island (Critically Endangered) and; e Cornufer citrinospilus from New Britain (Vulnerable). Photographs F. Kraus (b–d), S. Richards (e).Full size imageAll Critically Endangered and Endangered—and most of the Vulnerable—species were listed because of their small extent of occurrence and on-going decline in habitat area and/or quality (criteria B1ab(iii)) (Supplementary Table 6). The key threatening processes were typically forest disturbance or loss due to conversion to plantations or gardens, repeated burning, or mining (Fig. 4b–c). Only two insular species with very localised montane distributions were considered threatened by climatic disturbance and/or climate change alone (Cornufer citrinospilus and Oreophryne ezra) (Fig. 4d–e). No species were currently declining from pathogens, and in particular Batrachochytrium dendrobatidis (Bd), which remains undetected in Melaneisa51. However, the introduction and establishment of Bd has been identified as a severe threat for well over one hundred taxa52, especially for montane pelodryadid treefrogs, a group that has been devastated by this disease in parts of Australia.Although much of New Guinea has historically been considered a ‘wilderness area’ with comparatively little human impact53, the distributions of threatened taxa also highlight areas of conservation concern wherein range-restricted (often single-island endemic) taxa overlap with extensive and increasing anthropogenic impacts (Fig. 5a–b). Nearly half the species identified as threatened (13) are restricted to a recently delineated dramatic centre of herpetofaunal endemism in the Milne Bay Region at the eastern tip of Papua New Guinea21. Three clusters of small-range endemics in this region (all documented in the last two decades) present immediate conservation issues. The first is Mount Simpson, where six microhylids (four named, two awaiting description) with highly restricted ranges are threatened by habitat loss, especially repeated burning and associated conversion of forest to grassland (Fig. 4b). The second is Woodlark Island, where the status of seven endemic microhylids (six named, one undescribed) is likely to worsen rapidly if current, approved proposals to convert large areas of primary forest to oil-palm plantation and/or gold mines proceed21. Finally, Misima Island is home to four endemic microhylids (two considered threatened) with ranges that overlap areas disturbed by mining and forest loss21. Other regions with multiple overlapping threatened taxa are the Adelbert Mountains in Morobe Province (two species), New Britain (two lowland species and one highland species), and Greater Bukida in the Solomon Islands (three lowland species). These clusters of narrow-range taxa highlight important—and in most cases largely overlooked—conservation priorities for Melanesian frogs (and likely other taxa as well21,54). The high percentage of Data Deficient species and low level of survey effort in many areas (especially Papua and West Papua Provinces, Indonesia) also raise the possibility that other threatened hotspots remain overlooked. One area of particular concern may be the island of Biak in Indonesia, which has lost much of its primary vegetation but is home to at least three endemic frogs (one Data Deficient, two Least Concern).Fig. 5: The distribution of threatened frogs in Melanesia.a The estimated distribution of all 31 Melanesian frog species considered Critically Endangered, Endangered or Vulnerable at the end of 2019. Distributional areas are not colour coded by the number of threatened taxa. b Close up of the Milne Bay endemism hotspot. Distributional areas are colour coded by number of taxa, with darker tones indicating more taxa. In both a and b upland areas or islands where the distributions of two or more threatened species overlap are labelled and the number of threatened taxa are indicated in parentheses. Background maps uses the Shuttle Radar Topography Mission (SRTM) 30-meter digital elevation model, accessed from USGS Earth Explorer (https://earthexplorer.usgs.gov/).Full size imageUnderstanding and conserving a megadiverse biotaThe Melanesian flora and frog fauna are both now shown to be megadiverse and highly endemic, yet both also remain poorly known with large areas under-surveyed. An updated comprehensive assessment of threats and taxonomic trends across the frog fauna presented here further highlights that the biota of Melanesia remains relatively intact and less threatened when compared to other biodiverse insular regions. However, a large proportion of the fauna remains Data Deficient or undescribed, and key hotspots of endemism have been overlooked and are increasingly threatened. In both plants and anurans much scientific knowledge of Melanesia’s biota has also been contributed by a relatively small number of productive, but later-career researchers based outside of Melanesia7.Further documenting and conserving the exceptional diversity of Melanesia presents a suite of challenges and opportunities. Recommendations to enable improved documentation of plant megadiversity in Melanesia7 centre around training, capacity-building and support for taxonomy in Melanesia and globally, improving access to specimen collections and diagnostic resources, and ongoing support for survey and collecting within Melanesia. These recommendations apply equally to amphibians. However, addressing these challenges is tempered by the limited career opportunities available to ecologists and taxonomists (both in developed, but especially in developing countries), the variable quality of scientific infrastructure that exists across the region, and the high cost of doing fieldwork in remote areas with limited logistical infrastructure. In the context of these challenges, we hereby focus on suggesting some short-term key priorities and opportunities to build capacity for understanding and conserving frog biodiversity in Melanesia.First, over the last twenty years opportunities to employ Melanesian nationals in survey, monitoring and outreach work have been (and will continue to be) generated predominantly by NGOs, universities and large-scale extractive projects, for example through recent work in the gas fields of the Papua New Guinea Highlands49. While there are diverse perspectives on extractive industries, monitoring and survey work associated with large development projects are a key source of funds to provide training to enable Melanesians to undertake biodiversity work within the region. A key driver of this is strong environmental legislation required by some governments and major lending agencies, in particular the International Finance Corporation under Performance Standard 655. These requirements need to be maintained, enforced and, where possible, exceeded.To further support fieldwork by national scientists there is a need for more readily accessible identification resources for Melanesian researchers, land-owners and managers. An up-to-date comprehensive identification guide to the frog fauna of the whole region would assist and promote taxonomic, ecological and conservation research. However, for many Melanesians, small, regionally focused guides are more usable. These have already been produced for several areas (Supplementary References), providing a model that can be updated and transferred to other regions. Mobile phones are widely used throughout Melanesia, so app- and online-based identification resources may become increasingly accessible. Smartphone-friendly citizen science platforms like iNaturalist56 or even Facebook groups57 also provide potentially powerful resources through which locally collected data can be captured, vetted and disseminated, although their use is currently limited in Melanesia due to patchy internet coverage in many areas. Working with and supporting people from Melanesia to explore and increase the use of these resources could help to ensure longer-term preservation and accessibility of species records and associated data.The latest IUCN assessment for Melanesian frogs also highlights how taxonomic and conservation knowledge is accumulating rapidly. The key geographic areas of threat identified in our study were largely invisible to assessments made less than two decades ago (in 2004) both because the relevant taxonomic work had not been done, and because the situation in Melanesia is changing rapidly. To keep track of these rapid changes it is critical for workers in the region to work together to synthesise and collate new taxonomic, distributional and conservation data. Indeed, since the 2019 IUCN assessment over 20 additional species of Melanesian frogs have been described, and their conservation status should be assessed as a matter of urgency. Preliminary conservation assessments against IUCN criteria are increasingly being included in descriptions, and this trend should be supported and encouraged. More Melanesian nationals need to be involved in conservation assessment processes. Updated comprehensive conservation assessments of other vertebrate groups will also identify complementarity of conservation priorities among taxa in the Melanesian region.Patterns of distribution and threat suggest some geographic priority areas for documenting the diversity of amphibians (and potentially other low-vagility taxa) in Melanesia. First, work in eastern New Guinea has allowed the delineation of geographically localised clusters of threatened taxa that have until now gone unnoticed, perhaps in part because of the designation of much of Melanesia as a sparsely populated and comparatively undisturbed ‘wilderness’ area21. Most threatened frog taxa in these regions are associated with small islands or isolated ‘sky island’ mountains. The degree to which other taxa show endemism in these areas is poorly known. The biotas of potentially comparable islands in Indonesia such as the Raja Ampat Islands, Geelvink Bay and southern Maluku, also remain poorly known, suggesting additional priority areas for survey, taxonomic investigation and conservation assessment. Second, mid-elevation areas show highest alpha diversity, but large areas of this habitat, especially along the northern slopes of the Central Cordillera, remain poorly surveyed. The frog pathogen Bd has devastated montane communities of two Australian frog families that also occur in montane New Guinea (Myobatrachidae and Pelodryadidae)52. In the unfortunate event that Bd colonised New Guinea a wave of rapid declines and extinctions would likely follow52, so a strong baseline of information on montane species diversity, distributions and population status is critical for detecting these impacts. More

Our results show that GoM BFT larval survey samples can provide the crucial mark events for eventual CKMR estimates of adult abundance. The adult parents marked by larval samples can be directly recaptured in the fishery many years later as POPs, and indirectly through their progeny in future samples of larvae, as evidenced by the two cross-cohort HSPs (XHSPs) recovered in this study, which imply that a parent survived and spawned in the GoM in consecutive years. As more cohorts are sampled in future, the growing number of XHSPs could be used to estimate average adult survival rates, in addition to helping with the estimation of adult abundance31, as is now done for southern blue tuna40.There is a modest level of sibship within our 2016 samples, and a high level (involving over half the samples) in 2017, but it turns out not to be high enough to cause serious problems for POP-based CKMR. High sibship per se does not lead to bias in CKMR by virtue of the statistical construction of the estimate, but it does increase variance, which can be summarized through a reduction in effective sample size. In a POP-based CKMR model, our effective sample size would be about 75% of nominal for the two years combined, or 66% of nominal for the targeted sampling of 2017. Since it is actually the product of adult and juvenile sample sizes which drives precision in CKMR14, one way to think about the 75% is that we will need about 33% more adult samples to achieve a given precision on abundance estimates than if we had somehow been able to collect the same number of “independent” juvenile samples (i.e. without oversampling siblings). That increase is appreciable but entirely achievable; for WBFT, it is logistically much easier to collect more feeding-ground adult samples than to collect more larvae, and at present there is no known practical way to collect large numbers of older, more dispersed, and thus more independent, juvenile western origin bluefin tuna (WBFT).This study was motivated by the concern that sibship might be a serious impediment to use of WBFT larvae for CKMR. High levels of sibship have been found in larval collections for other taxa despite a pelagic larval phase, suggesting that abiotic factors can impede random mixing of larvae after a spawning event41. Our larval samples were only a few days old (4–11) and thus had little time to disperse since fertilization; our concern beforehand was that each tow might sample the offspring of a very small number of adults (one spawning group in one night), and in 2017 that repeatedly towing the same water mass might simply be resampling the same “family”. In practice, though, the cumulative effect was limited. Samples were not dominated by progeny from just a few adults; the maximum DPG size (i.e., number of offspring from any one adult) was 5, which is under 2% of the larval sample size. There are several possible reasons for this finding. First, plankton sample tows are typically standardized to a ten-minute duration, covering on average about 0.3 nautical miles. Based on continuous plankton cameras42, each tow is likely to tow through multiple patches of zooplankton, and therefore potentially multiple patches of BFT larvae. Second, spawning aggregations of BFT may contain many adults. For example, on the spawning grounds near the Balearic Islands in the Mediterranean, purse seine fisheries target spawning fish and individual net sets routinely capture upwards of 500 mature individuals43. These numbers suggest that BFT spawner aggregations can be quite large, although the number of individuals that contribute gametes to a single spawning event may be lower. The results of this study pose intriguing scenarios for understanding BFT larval ecology and spawning behavior, which could be explored with larger sample sizes paired with data on oceanographic conditions, direct observation of spawning aggregations, and modeling to compare observed and predicted dispersal. The results of this study are based on just two years of sampling, and numerous practical and theoretical challenges remain to fully understand BFT reproduction in the GoM.Our sibship impact calculations assume use of an unmodified adult-size-based CKMR POP model, where each juvenile is compared to each adult taking into account the latter’s size (e.g.,14). That will give unbiased estimates, which we regard as essential in a CKMR model. However, for WBFT the estimates are not fully statistically efficient, in that some adults receive more statistical weight than others because they are marked more often (by having a large DPG), and thus variance might not be the lowest achievable. Modifying the model to fix that would be simple in a “cartoon” CKMR setting where all adults are identical (e.g., Fig. 1 of14), simply by first condensing each DPG to a single representative, then only using those representatives (rather than all the larvae) in POP comparisons. Each marked parent then receives the same weight, giving maximum efficiency. For the cartoon, this condensed-DPG model still gives an unbiased estimate of abundance, because each DPG has one parent of given sex, and the chance of any sampled cartoon adult of that sex being that parent is 1/N. The DPG-condensed effective sample size is simply half the number of distinct parents, which would be a little larger than the effective sample sizes for the unmodified model shown in Table 3; e.g., in 2017, 504/2 = 252 versus 209. However, no such straightforward improvement is available for an adult-size-based CKMR model such as is needed for WABFT. Using condensed DPGs directly would bias the juvenile sampling against larger more-fecund adults, whose DPGs will tend on average to be larger and thus to experience disproportionate condensation. Those adults would be marked less often by the DPG-condensed juveniles than the model assumes, violating the basic requirements for unbiased CKMR in14. A more sophisticated model might be able to combine unbiasedness with higher efficiency but, since the unmodified adult-size-based POP model that we expect to use is unbiased and only mildly inefficient (at worst 209/252 = 83% efficient, in 2017) there seems no particular need for extra complications at present. However, that may not hold true if we eventually move to a POP + XHSP model, where the impact on unmodified CKMR variance is worse (though there is still no bias, for the same reason as with POPs). Intuitively, the biggest impact that a DPG of size 5 can have in a POP model is to suddenly raise the number of POPs by 5 if its parent happens to be sampled; within a useful total of, say, 75 POPs, the influence is not that large. But if two DPGs both of size 5 in different cohorts happen to share a parent, then the total of XHSPs suddenly jumps by 25— likely a substantial proportion of total XHSPs. Supplementary Material B also includes effective sample size formulae for a simplified XHSP-only model, which demonstrate the increased impact of within-cohort sibship; for our WBFT samples, it turns out that the XHSP-effective size is slightly lower for the targeted 2017 samples (110) than for the 2016 samples (130), unlike the POP-only effective size. Dropping from a maximum theoretical effective sample size of 252 (half the number of DPGs) down to 110 would be rather inefficient and would increase the number of years of sampling required to yield a useful XHSP dataset. This motivates developing a modified POP + XHSP model that retains unbiasedness without sacrificing too much efficiency. In principle, that can be done by condensing each DPG but then conditioning its comparison probabilities on the DPG’s original size, in accordance with the framework in14. This is a topic for subsequent research, and the results will inform future sampling strategy decisions for WBFT.One potential difficulty for western BFT CKMR might occur if a substantial proportion of animals reaching maturity are the offspring of “Western” (in genetic terms) adults who persistently spawn in the western North Atlantic but outside the GoM. However, as long as the adults marked by GoM larvae are well mixed at the time of sampling with any western adults that do spawn outside of the GoM, the total POP-based population estimate of genetically-western BFT from CKMR will remain unbiased. Given evidence from tagging of widespread adult movements within the western North Atlantic2, good mixing in the sampled feeding grounds seems likely; so, even if successful non-GoM western BFT spawning really is commonplace, there should not be a problem with relying on GoM larvae for at least the POP component of CKMR14.Studies of fish early life history have long been considered to have great potential to provide novel insight into the unique population dynamics of fishes44,45,46. Sampling efforts aimed at estimating fish recruitment dynamics have spawned a diversity of larval survey programs. Examples of these long-term programs include the California Cooperative Oceanic Fisheries Investigations, International Council for the Exploration of the Sea (ICES) surveys in the North Atlantic and adjacent areas, Southeast Monitoring and Assessment Program (SEAMAP) in the GoM, Ecosystem Monitoring (EcoMon) in the Northeast U.S., and numerous others, many of which provide indices of larval abundance widely used in fisheries and ecosystem assessments. Yet, as a result of the inherent patchiness of larvae42, sampling variability, and highly variable density dependent mortality45, fisheries scientists have often struggled to determine how larval surveys relate to the adult fish populations. Inclusion of estimates of sibship among larvae collected in surveys could refine estimates of adult spawning stock biomass estimated from these surveys.The results of this study also represent products of decades of work and coordination in obtaining high-quality DNA from larval specimens. Key steps to successful genotyping of larvae include ensuring that larvae are preserved, sorted, and handled in 95% non-denatured ethanol. In addition, strict instrument cleaning protocols must be followed, and stomachs should be removed or avoided (this study used larval tails and, when possible, eyes to avoid cross contamination of prey contents, including possible congeners and other BFT individuals). Exposure to hot lamps during the sorting and dissection processes should also be minimized to ensure that DNA quality is sufficiently high for genotyping-by-sequencing. Although the tissues available for genetic analysis were limited by the needs of other experiments that required BFT tissues, otoliths, gut contents, and other information from the same larvae, we were able to successfully genotype most larvae greater than 6 mm SL and identify thousands of informative SNPs. The lower size limit of larvae could likely be decreased if whole specimens were available for genotyping, although the use of younger larvae could increase the incidence of sibship.In summary, while we observed both FSPs and HSPs in larval collections, with elevated sibship overall and with siblings being more prevalent within tows and in nearby tows, the level of sibship was sufficiently low that collections of GoM BFT larvae can still provide the critical genetic mark of parental genotypes required for CKMR. Our results demonstrate a crucial proof of concept and are the first step towards an operational CKMR modelling estimate of spawning stock abundance for western BFT. More