Ecology

1.
SAUP. Sea Around Us. http://www.seaaroundus.org/data/ (2020).
2.
Coll, M., Navarro, J., Olson, R. J. & Christensen, V. Assessing the trophic position and ecological role of squids in marine ecosystems by means of food-web models. Deep Sea Res. Part II Top. Stud. Oceanogr. 95, 21–36 (2013).
ADS  Article  Google Scholar 

3.
Hastie, L. et al. Cephalopods in the north-eastern Atlantic: Species, biogeography, ecology, exploitation and conservation. In Oceanography and Marine Biology (eds. Gibson, R., Atkinson, R. & Gordon, J.) vol. 20092725, 111–190 (CRC Press, Boca Raton, 2009).

4.
Piatkowski, U. & Pierce, G. J. Impact of cephalopods in the food chain and their interaction with the environment and fisheries: An overview. Fish. Res. 6, 5–10 (2001).
Article  Google Scholar 

5.
Pierce, G. J. et al. A review of cephalopod–environment interactions in European Seas. Hydrobiologia 612, 49–70 (2008).
Article  Google Scholar 

6.
André, J., Haddon, M. & Pecl, G. T. Modelling climate-change-induced nonlinear thresholds in cephalopod population dynamics. Glob. Change Biol. 16, 2866–2875 (2010).
ADS  Article  Google Scholar 

7.
Jereb, P. et al. Cephalopod biology and fisheries in Europe: II. Species Accounts. ICES Cooper. Res. Rep. 325, 1–360 (2015).
Google Scholar 

8.
Sims, D. W., Genner, M. J., Southward, A. J. & Hawkins, S. J. Timing of squid migration reflects North Atlantic climate variability. Proc. R. Soc. Lond. Ser. B Biol. Sci. 268, 2607–2611 (2001).

9.
Dorey, N. et al. Ocean acidification and temperature rise: Effects on calcification during early development of the cuttlefish Sepia officinalis. Mar. Biol. 160, 2007–2022 (2013).
CAS  Article  Google Scholar 

10.
Rodhouse, P. G. K. et al. Environmental effects on cephalopod population dynamics. In Advances in Marine Biology vol. 67, 99–233 (Elsevier, Amsterdam, 2014).

11.
Millar, R. J. et al. Emission budgets and pathways consistent with limiting warming to 1.5 °C. Nat. Geosci. 10, 741–747 (2017).
ADS  CAS  Article  Google Scholar 

12.
Otto, F. E. L., Frame, D. J., Otto, A. & Allen, M. R. Embracing uncertainty in climate change policy. Nat. Clim. Change 5, 917–920 (2015).
ADS  Article  Google Scholar 

13.
Meinshausen, M. et al. The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Clim. Change 109, 213–241 (2011).
ADS  CAS  Article  Google Scholar 

14.
van Vuuren, D. P. et al. The representative concentration pathways: An overview. Clim. Change 109, 5–31 (2011).
ADS  Article  Google Scholar 

15.
Gissi, E. et al. A review of the combined effects of climate change and other local human stressors on the marine environment. Sci. Total Environ. 755, 142564 (2021).
ADS  CAS  PubMed  Article  Google Scholar 

16.
Beaugrand, G. et al. Prediction of unprecedented biological shifts in the global ocean. Nat. Clim. Change 9, 237–243 (2019).
ADS  Article  Google Scholar 

17.
Jorda, G. et al. Ocean warming compresses the three-dimensional habitat of marine life. Nat. Ecol. Evol. 4, 109–114 (2020).
PubMed  Article  Google Scholar 

18.
Vidal, E. A. G., DiMarco, F. P., Wormuth, J. H. & Lee, P. G. Influence of temperature and food availability on survival, growth and yolk utilization in hatchling squid. Bull. Mar. Sci. 71, 915–931 (2002).
Google Scholar 

19.
Doubleday, Z. A. et al. Global proliferation of cephalopods. Curr. Biol. 26, R406–R407 (2016).
CAS  PubMed  Article  Google Scholar 

20.
van der Kooij, J., Engelhard, G. H. & Righton, D. A. Climate change and squid range expansion in the North Sea. J. Biogeogr. 43, 2285–2298 (2016).
Article  Google Scholar 

21.
Jin, Y., Jin, X., Gorfine, H., Wu, Q. & Shan, X. Modeling the oceanographic impacts on the spatial distribution of common cephalopods during autumn in the yellow sea. Front. Mar. Sci. 7, (2020).

22.
Pang, Y. et al. Variability of coastal cephalopods in overexploited China Seas under climate change with implications on fisheries management. Fish. Res. 208, 22–33 (2018).
Article  Google Scholar 

23.
Le Marchand, M. et al. Climate change in the Bay of Biscay: Changes in spatial biodiversity patterns could be driven by the arrivals of southern species. Mar. Ecol. Prog. Ser. 647, 17–31 (2020).
ADS  Article  Google Scholar 

24.
Lima, F. D., Ángeles-González, L. E., Leite, T. S. & Lima, S. M. Q. Global climate changes over time shape the environmental niche distribution of Octopus insularis in the Atlantic Ocean. Mar. Ecol. Prog. Ser. 652, 111–121 (2020).
ADS  Article  Google Scholar 

25.
Xavier, J. C., Peck, L. S., Fretwell, P. & Turner, J. Climate change and polar range expansions: Could cuttlefish cross the Arctic?. Mar. Biol. 163, 78 (2016).
Article  Google Scholar 

26.
Selig, E. R. et al. Mapping global human dependence on marine ecosystems. Conserv. Lett. 12, e12617 (2019).
Article  Google Scholar 

27.
Blasiak, R. et al. Climate change and marine fisheries: Least developed countries top global index of vulnerability. PLoS ONE 12, e0179632 (2017).
PubMed  PubMed Central  Article  CAS  Google Scholar 

28.
FAO. The State of Mediterranean and Black Sea Fisheries. (General Fisheries Commission for the Mediterranean, 2016).

29.
Lam, V. W. Y., Cheung, W. W. L., Reygondeau, G. & Sumaila, U. R. Projected change in global fisheries revenues under climate change. Sci. Rep. 6, 32607 (2016).
ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

30.
Badjeck, M.-C., Perry, A., Renn, S., Brown, D. & Poulain, F. The vulnerability of fishing-dependent economies to disasters. FAO Fish. Aquac. Circ. 1081, 1–19 (2013).
Google Scholar 

31.
Allison, E. H. et al. Vulnerability of national economies to the impacts of climate change on fisheries. Fish Fish. 10, 173–196 (2009).
Article  Google Scholar 

32.
Adloff, F. et al. Mediterranean Sea response to climate change in an ensemble of twenty first century scenarios. Clim. Dyn. 45, 2775–2802 (2015).
Article  Google Scholar 

33.
Alexander, M. A. et al. Projected sea surface temperatures over the 21st century: Changes in the mean, variability and extremes for large marine ecosystem regions of Northern Oceans. Elementa Sci. Anthropocene 6, 9 (2018).
Article  Google Scholar 

34.
Gaines, S. D. et al. Improved fisheries management could offset many negative effects of climate change. Sci. Adv. 4, eaao1378 (2018).
ADS  PubMed  PubMed Central  Article  Google Scholar 

35.
Pierce, G. J. et al. Status and trends of European cephalopod stocks. In ASC 2019 ICES Conference, Gothenburg, Sweden 1 (2019).

36.
Hutchinson, G. E. Concluding remarks. Cold Spring Harb. Symp. Quant. Biol. 22, 415–427 (1957).
Article  Google Scholar 

37.
Hutchinson, G. E. An Introduction to Population Ecology (Yale University Press, New Haven, 1978).
Google Scholar 

38.
Peterson, A. & Soberón, J. Species distribution modeling and ecological niche modeling: Getting the concepts right. Natureza e Conservação 10, 1–6 (2012).
Article  Google Scholar 

39.
Colwell, R. K. & Rangel, T. F. Hutchinson’s duality: The once and future niche. Proc. Natl. Acad. Sci. 106, 19651–19658 (2009).
ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

40.
Buisson, L., Thuiller, W., Casajus, N., Lek, S. & Grenouillet, G. Uncertainty in ensemble forecasting of species distribution. Glob. Change Biol. 16, 1145–1157 (2010).
ADS  Article  Google Scholar 

41.
Araújo, M. B. & New, M. Ensemble forecasting of species distributions. Trends Ecol. Evol. 22, 42–47 (2007).
PubMed  Article  Google Scholar 

42.
Hao, T., Elith, J., Guillera-Arroita, G. & Lahoz-Monfort, J. J. A review of evidence about use and performance of species distribution modelling ensembles like BIOMOD. Divers. Distrib. https://doi.org/10.1111/ddi.12892 (2019).
Article  Google Scholar 

43.
Goberville, E., Beaugrand, G., Hautekèete, N.-C., Piquot, Y. & Luczak, C. Uncertainties in the projection of species distributions related to general circulation models. Ecol. Evol. 5, 1100–1116 (2015).
PubMed  PubMed Central  Article  Google Scholar 

44.
Leroy, B. et al. Forecasted climate and land use changes, and protected areas: The contrasting case of spiders. Divers. Distrib. 20, 686–697 (2014).
Article  Google Scholar 

45.
Schickele, A. et al. Modelling European small pelagic fish distribution: Methodological insights. Ecol. Model. 416, 108902 (2020).
Article  Google Scholar 

46.
Araújo, M. B. et al. Standards for distribution models in biodiversity assessments. Sci. Adv. 5, eaat4858 (2019).
ADS  PubMed  PubMed Central  Article  Google Scholar 

47.
Barbet-Massin, M., Thuiller, W. & Jiguet, F. How much do we overestimate future local extinction rates when restricting the range of occurrence data in climate suitability models?. Ecography 33, 878–886 (2010).
Article  Google Scholar 

48.
Beaugrand, G., Luczak, C., Goberville, E. & Kirby, R. Marine biodiversity and the chessboard of life. PLoS ONE 13, e0194006 (2018).
PubMed  PubMed Central  Article  CAS  Google Scholar 

49.
Støa, B., Halvorsen, R., Mazzoni, S. & Gusarov, V. I. Sampling bias in presence-only data used for species distribution modelling: Theory and methods for detecting sample bias and its effects on models. Sommerfeltia 38, 1–53 (2018).
Article  Google Scholar 

50.
Dufresne, J.-L. et al. Climate change projections using the IPSL-CM5 Earth System Model: from CMIP3 to CMIP5. Clim. Dyn. 40, 2123–2165 (2013).
Article  Google Scholar 

51.
Voldoire, A. et al. The CNRM-CM5.1 global climate model: Description and basic evaluation. Clim. Dyn. 40, 2091–2121 (2013).
Article  Google Scholar 

52.
Hourdin, F. et al. Impact of the LMDZ atmospheric grid configuration on the climate and sensitivity of the IPSL-CM5A coupled model. Clim. Dyn. 40, 2167–2192 (2013).
Article  Google Scholar 

53.
Friedlingstein, P. et al. Uncertainties in CMIP5 climate projections due to carbon cycle feedbacks. J. Clim. 27, 511–526 (2013).
ADS  Article  Google Scholar 

54.
Wiens, J. A., Stralberg, D., Jongsomjit, D., Howell, C. A. & Snyder, M. A. Niches, models, and climate change: Assessing the assumptions and uncertainties. PNAS 106, 19729–19736 (2009).
ADS  CAS  PubMed  Article  Google Scholar 

55.
Martinez-Meyer, E. Climate change and biodiversity: Some considerations in forecasting shifts in species’ potential distributions. Biodivers. Inform. 2, 42–55 (2005).
Article  Google Scholar 

56.
Levitus, S. Climatological atlas of the world ocean. Eos Trans. Am. Geophys. Union 64, 962–963 (2011).
ADS  Article  Google Scholar 

57.
Cabanes, C. et al. The CORA dataset: Validation and diagnostics of in-situ ocean temperature and salinity measurements. Ocean Sci. 9, 1–18 (2013).
ADS  Article  Google Scholar 

58.
Giorgetta, M. A. et al. Climate and carbon cycle changes from 1850 to 2100 in MPI-ESM simulations for the Coupled Model Intercomparison Project phase 5: Climate Changes in MPI-ESM. J. Adv. Model. Earth Syst. 5, 572–597 (2013).
ADS  Article  Google Scholar 

59.
Stevens, B. et al. Atmospheric component of the MPI-M Earth System Model: ECHAM6. J. Adv. Model. Earth Syst. 5, 146–172 (2013).
ADS  Article  Google Scholar 

60.
Jones, C. D. et al. The HadGEM2-ES implementation of CMIP5 centennial simulations. Geosci. Model Dev. Discuss. 4, 689–763 (2011).
ADS  Google Scholar 

61.
Schmidt, G. A. et al. Configuration and assessment of the GISS ModelE2 contributions to the CMIP5 archive: GISS MODEL-E2 CMIP5 SIMULATIONS. J. Adv. Model. Earth Syst. 6, 141–184 (2014).
ADS  Article  Google Scholar 

62.
Beaugrand, G., Lenoir, S., Ibañez, F. & Manté, C. A new model to assess the probability of occurrence of a species, based on presence-only data. Mar. Ecol. Prog. Ser. 424, 175–190 (2011).
ADS  Article  Google Scholar 

63.
Raybaud, V., Bacha, M., Amara, R. & Beaugrand, G. Forecasting climate-driven changes in the geographical range of the European anchovy (Engraulis encrasicolus). ICES J. Mar. Sci. 74, 1288–1299 (2017).
Article  Google Scholar 

64.
Thuiller, W., Lafourcade, B., Engler, R. & Araújo, M. B. BIOMOD—A platform for ensemble forecasting of species distributions. Ecography 32, 369–373 (2009).
Article  Google Scholar 

65.
Thuiller, W., Georges, D., Engler, R. & Breiner, F. Ensemble Platform for Species Distribution Modelling. (2016).

66.
Dormann, C. F. et al. Collinearity: A review of methods to deal with it and a simulation study evaluating their performance. Ecography 36, 27–46 (2013).
Article  Google Scholar 

67.
Lenoir, J. et al. Species better track climate warming in the oceans than on land. Nat. Ecol. Evol. https://doi.org/10.1038/s41559-020-1198-2 (2020).
Article  PubMed  Google Scholar 

68.
Smith, W. H. F. & Sandwell, D. T. Global sea floor topography from satellite altimetry and ship depth soundings. Science 277, 1956–1962 (1997).
CAS  Article  Google Scholar 

69.
NASA Goddard Space Flight Center, Ocean Ecology Laboratory, Ocean Biology Processing Group. Distance to the nearest coast. https://oceancolor.gsfc.nasa.gov/docs/distfromcoast/ (01/03/2018) (2009).

70.
Hattab, T. et al. Towards a better understanding of potential impacts of climate change on marine species distribution: A multiscale modelling approach. Glob. Ecol. Biogeogr. 23, 1417–1429 (2014).
Article  Google Scholar 

71.
Varela, S., Anderson, R. P., García-Valdés, R. & Fernández-González, F. Environmental filters reduce the effects of sampling bias and improve predictions of ecological niche models. Ecography 37, 1084–1091 (2014).
Google Scholar 

72.
Ben Rais Lasram, F. et al. An open-source framework to model present and future marine species distributions at local scale. Ecol. Inform. 59, 101130 (2020).
Article  Google Scholar 

73.
Montgomery, D. C. Design and Analysis of Experiments (Wiley, Hoboken, 2005).
Google Scholar 

74.
Getz, W. M. & Wilmers, C. C. A local nearest-neighbor convex-hull construction of home ranges and utilization distributions. Ecography 27, 489–505 (2006).
Article  Google Scholar 

75.
Cornwell, W. K., Schwilk, D. W. & Ackerly, D. D. A trait-based test for habitat filtering: Convex hull volume. Ecology 87, 1465–1471 (2004).
Article  Google Scholar 

76.
Hirzel, A. H., Le Lay, G., Helfer, V., Randin, C. & Guisan, A. Evaluating the ability of habitat suitability models to predict species presences. Ecol. Model. 199, 142–152 (2006).
Article  Google Scholar 

77.
Leroy, B. et al. Without quality presence–absence data, discrimination metrics such as TSS can be misleading measures of model performance. J. Biogeogr. 45, 1994–2002 (2018).
Article  Google Scholar 

78.
Faillettaz, R., Beaugrand, G., Goberville, E. & Kirby, R. R. Atlantic Multidecadal Oscillations drive the basin-scale distribution of Atlantic bluefin tuna. Sci. Adv. 5, eaar6993 (2019).
ADS  PubMed  PubMed Central  Article  Google Scholar 

79.
Elith, J., Ferrier, S., Huettmann, F. & Leathwick, J. The evaluation strip: A new and robust method for plotting predicted responses from species distribution models. Ecol. Model. 186, 280–289 (2005).
Article  Google Scholar 

80.
VanDerWal, J. et al. Focus on poleward shifts in species’ distribution underestimates the fingerprint of climate change. Nat. Clim. Change 3, 239–243 (2013).
ADS  Article  Google Scholar 

81.
Taylor, K. E. Summarizing multiple aspects of model performance in a single diagram. J. Geophys. Res. Atmos. 106, 7183–7192 (2001).
ADS  Article  Google Scholar 

82.
Cristofari, R. et al. Climate-driven range shifts of the king penguin in a fragmented ecosystem. Nat. Clim. Change 8, 245–251 (2018).
ADS  Article  Google Scholar 

83.
Péron, C., Weimerskirch, H. & Bost, C.-A. Projected poleward shift of king penguins’ (Aptenodytes patagonicus) foraging range at the Crozet Islands, southern Indian Ocean. Proc. Biol. Sci. 279, 2515–2523 (2012).
PubMed  PubMed Central  Article  Google Scholar 

84.
Bloor, I. S. M., Attrill, M. J. & Jackson, E. L. Chapter One—A Review of the Factors Influencing Spawning, Early Life Stage Survival and Recruitment Variability in the Common Cuttlefish (Sepia officinalis). In Advances in Marine Biology (ed. Lesser, M.) vol. 65, 1–65 (Academic Press, Cambridge, 2013).

85.
Vidal, E. A. G., Roberts, M. J. & Martins, R. S. Yolk utilization, metabolism and growth in reared Loligo vulgaris reynaudii paralarvae. Aquat. Living Resour. 18, 385–393 (2005).
Article  Google Scholar 

86.
Bouchaud, O. Energy consumption of the cuttlefish Sepia officinalis L. (Mollusca: Cephalopoda) during embryonic development, preliminary results. Bull. Mar. Sci. 49, 333–340 (1991).
Google Scholar 

87.
Laptikhovsky, V. Latitudinal and bathymetric trends in egg size variation: A new look at Thorson’s and Rass’s rules. Mar. Ecol. 27, 7–14 (2006).
ADS  Article  Google Scholar 

88.
Hengl, T., Sierdsema, H., Radović, A. & Dilo, A. Spatial prediction of species’ distributions from occurrence-only records: Combining point pattern analysis, ENFA and regression-kriging. Ecol. Model. 220, 3499–3511 (2009).
Article  Google Scholar 

89.
Clarke, M. R. The role of cephalopods in the world’s oceans: general conclusions and the future. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 351, 1105–1112 (1996).
ADS  Article  Google Scholar 

90.
Marmion, M., Parviainen, M., Luoto, M., Heikkinen, R. K. & Thuiller, W. Evaluation of consensus methods in predictive species distribution modelling. Divers. Distrib. 15, 59–69 (2009).
Article  Google Scholar 

91.
Kissling, W. D. et al. Towards novel approaches to modelling biotic interactions in multispecies assemblages at large spatial extents: Modelling multispecies interactions. J. Biogeogr. 39, 2163–2178 (2012).
Article  Google Scholar 

92.
Clark, J. S., Gelfand, A. E., Woodall, C. & Zhu, K. More than the sum of the parts: Forest climate response from joint species distribution models. Ecol. Appl. 24, 990–999 (2014).
PubMed  Article  Google Scholar 

93.
Harris, D. J. Generating realistic assemblages with a joint species distribution model. Methods Ecol. Evol. 6, 465–473 (2015).
Article  Google Scholar 

94.
Nogués-Bravo, D. Predicting the past distribution of species climatic niches. Glob. Ecol. Biogeogr. 18, 521–531 (2009).
Article  Google Scholar 

95.
Lee, Q., Thorson, J. T., Gertseva, V. V. & Punt, A. E. The benefits and risks of incorporating climate-driven growth variation into stock assessment models, with application to Splitnose Rockfish (Sebastes diploproa). ICES J. Mar. Sci. 75, 245–256 (2018).
Article  Google Scholar 

96.
Colléter, M., Gascuel, D., Ecoutin, J.-M. & Tito de Morais, L. Modelling trophic flows in ecosystems to assess the efficiency of marine protected area (MPA), a case study on the coast of Sénégal. Ecol. Model. 232, 1–13 (2012).
Article  Google Scholar 

97.
Allen, K. R. Relation between production and biomass. J. Fish. Res. Board Can. 28, 1573–1581 (1971).
Article  Google Scholar 

98.
FAO. Review of the state of world marine fishery resources. FAO Fish. Aquac. Tech. Pap. 334 (2011).

99.
Cheung, W. W. L. et al. Transform high seas management to build climate resilience in marine seafood supply. Fish Fish. 18, 254–263 (2016).
Article  Google Scholar 

100.
Sumaila, U. R., Cheung, W. W. L., Lam, V. W. Y., Pauly, D. & Herrick, S. Climate change impacts on the biophysics and economics of world fisheries. Nat. Clim. Change 1, 449–456 (2011).
ADS  Article  Google Scholar 

101.
Barange, M. et al. Impacts of climate change on marine ecosystem production in societies dependent on fisheries. Nat. Clim. Change 4, 211–216 (2014).
ADS  Article  Google Scholar 

102.
Ojea, E., Lester, S. E. & Salgueiro-Otero, D. Adaptation of fishing communities to climate-driven shifts in target species. One Earth 2, 544–556 (2020).
Article  Google Scholar  More

Site description, chronology, and sampling
A detailed description of the site, coring methods, age-depth model reconstruction, and sampling strategy can be found in Alsos et al.41. Briefly, Lake Øvre Æråsvatnet is located on Andøya, Northern Norway (69.25579°N, 16.03517°E) (Fig. 1a, b). In 2013, two cores were collected from the deepest sediments, AND10 and AND11, which were stored at 4 °C prior to sampling. Macrofossil remains were dated, with those from AND10 all dating to within the LGM. For the longer core AND11, a Bayesian age-depth model was required to estimate the age of each layer41. In this study, we selected one sample of LGM sediments from each of the two cores. According to the Bayesian age-depth model, sample Andøya_LGM_B, from 1102 cm depth in AND11, was dated to a median age of 17,700 (range: 20,200–16,500) cal yr BP. The age of Andøya_LGM_A, from 938 cm depth in AND10, was estimated at 19,500 cal yr BP, based on the interpolated median date between two adjacent macrofossils (20 cm above: 19,940–18,980 cal yr BP, 30 cm below: 20,040–19,000 cal yr BP). As Andøya_LGM_A falls within the age range of Andøya_LGM_B, we consider the samples to be broadly contemporaneous.
Sampling, DNA extraction, library preparation, and sequencing
The two cores were subsampled at the selected layers under clean conditions, in a dedicated ancient DNA laboratory at The Arctic University Museum of Norway in Tromsø. We extracted DNA from 15 g of sediment following the Taberlet phosphate extraction protocol29 in the same laboratory. We shipped a 210 µL aliquot of each DNA extract to the ancient DNA dedicated laboratories at the Centre for GeoGenetics (University of Copenhagen, Denmark) for double-stranded DNA library construction. The extracts were concentrated to 80 µL using Amicon Ultra-15 30 kDa centrifugal filters (Merck Millipore, Darmstadt, Germany) and half of each extract (40 µL, totalling between 31.7 and 36.0 ng of DNA) was converted into Illumina-compatible libraries using established protocols10. Each library was dual-indexed via 12 cycles of PCR. The libraries were then purified using the AmpureBead protocol (Beckman Coulter, Indianapolis, IN, USA), adjusting the volume ratio to 1:1.8 library:AmpureBeads, and quantified using a BioAnalyzer (Agilent, Santa Clara, CA, USA). The indexed libraries were pooled equimolarly and sequenced on a lane of the Illumina HiSeq 2500 platform using 2 × 80 cycle paired-end chemistry.
Raw read filtering
For each sample, we merged and adapter-trimmed the paired-end reads with SeqPrep (https://github.com/jstjohn/SeqPrep/releases, v1.2) using default parameters. We only retained the resulting merged sequences, which were then filtered with the preprocess function of the SGA toolkit v0.10.15 (ref. 57) by the removal of those shorter than 35 bp or with a DUST complexity score > 1.
Metagenomic analysis of the sequence data
We first sought to obtain an overview of the taxonomic composition of the samples and therefore carried out a BLAST-based metagenomic analysis on the two filtered sequence datasets. To make the datasets more computationally manageable, we subsampled the first and last one-million sequences from the filtered dataset of each sample and analysed each separately. The data subsets were each identified against the NCBI nucleotide database (release 223) using the blastn function from the NCBI-BLAST+ suite v2.2.18+58 under default settings. For each sample, the results from the two subsets were checked for internal consistency, merged into one dataset, and loaded into MEGAN v6.12.3 (ref. 59). Analysis and visualization of the Last Common Ancestor (LCA) was carried out for the taxonomic profile using the following settings: min score = 35, max expected = 1.0E−5, min percent identity = 95%, top percent = 10%, min support percentage = 0.01, LCA = naive, min percent sequence to cover = 95%. We define sequences as the reads with BLAST hits assigned to taxa post-filtering, thus ignoring “unassigned” and “no hit” categories.
Alignment to reference genome panels
We mapped our filtered data against three different reference panels to help improve taxonomic identifications and provide insight into the sequence abundance of the identified taxa (Supplementary Data 2 and 3). The first reference panel consisted of 42 nuclear genomes that included genera expected from Northern Norway, exotic/implausible taxa for LGM Andøya, human, six Nannochloropsis species, and four strains of Mycobacterium. The inclusion of exotic taxa was to give an indication of the background spurious mapping rate, which can result from mappings to conserved parts of the genome and/or short and damaged ancient DNA molecules22,23. We included Nannochloropsis, Mycobacterium, and human genomes, due to their overrepresentation in the BLAST-based metagenomic analysis. The other two reference panels were based on either all 8486 mitochondrial or 2495 chloroplast genomes on NCBI GenBank (as of January 2018). The chloroplast dataset was augmented with 247 partial or complete chloroplast genomes generated by the PhyloNorway project60 for 2742 chloroplast genomes in total. The filtered data were mapped against each reference genome or organellar genome set individually using bowtie2 v2.3.4.1 (ref. 61) under default settings. The resulting bam files were processed with SAMtools v0.1.19 (ref. 62). We removed unmapped sequences with SAMtools view and collapsed PCR duplicate sequences with SAMtools rmdup.
For the nuclear reference panel, we reduced potential spurious or nonspecific sequence mappings by comparing the mapped sequences to both the aligned reference genome and the NCBI nucleotide database using NCBI-BLAST+, following the method used by Graham et al.9, as modified by Wang et al.11. The sequences were aligned using the following NCBI-BLAST+ settings: num_alignments = 100 and perc_identity = 90. Sequences were retained if they had better alignments, based on bit score, to reference genomes as compared to the NCBI nucleotide database. If a sequence had a better or equal match against the NCBI nucleotide database, it was removed, unless the LCA of the highest NCBI nucleotide bit score was from the same genus as the reference genome (based on the NCBI taxonID). To standardize the relative mapping frequencies to genomes of different size, we calculated the number of retained mapped sequences per Mb of genome sequence.
The sequences mapped against the chloroplast and mitochondrial reference panels were filtered and reported in a different manner than the nuclear genomes. First, to exclude any non-eukaryotic sequences, we used NCBI-BLAST+ to search sequence taxonomies and retained sequences if the LCA was, or was within, Eukaryota. Second, for the sequences that were retained, the LCA was calculated and reported in order to summarize the mapping results across the organelle datasets. LCAs were chosen as the reference sets are composed of multiple genera.
Within the Nannochloropsis nuclear reference alignments, the relative mapping frequency was highest for N. limnetica. In addition, the relative mapping frequency for other Nannochloropsis taxa was higher than those observed for the exotic taxa. This could represent the mapping of sequences that are conserved between Nannochloropsis genomes or suggest the presence of multiple Nannochloropsis taxa in a community sample. We therefore cross-compared mapped sequences to determine the number of uniquely mapped sequences per reference genome. First, we individually remapped the filtered data to six available Nannochloropsis nuclear genomes, the accession codes of which are provided in Supplementary Data 2. For each sample, we then calculated the number of sequences that uniquely mapped to, or overlapped, between each Nannochloropsis genome. We repeated the above analysis with six available chloroplast sequences (Supplementary Data 2) to get a comparable overlap estimation for the chloroplast genome.
Reconstruction of the Andøya Nannochloropsis community organellar palaeogenomes
To place the Andøya Nannochloropsis community taxon into a phylogenetic context, and provide suitable reference sequences for variant calling, we reconstructed environmental palaeogenomes for the Nannochloropsis mitochondria and chloroplast. First, the raw read data from both samples were combined into a single dataset and re-filtered with the SGA toolkit to remove sequences shorter than 35 bp, but retain low complexity sequences to assist in the reconstruction of low complexity regions in the organellar genomes. This re-filtered sequence dataset was used throughout the various steps for environmental palaeogenome reconstruction.
The re-filtered sequence data were mapped onto the N. limnetica reference chloroplast genome (NCBI GenBank accession: NC_022262.1) with bowtie2 using default settings. SAMtools was used to remove unmapped sequences and PCR duplicates, as above. We generated an initial consensus genome from the resulting bam file with BCFtools v1.9 (ref. 62), using the mpileup, call, filter, and consensus functions. For variable sites, we produced a majority-rule consensus using the –variants-only and –multiallelic-caller options, and for uncovered sites the reference genome base was called. The above steps were repeated until the consensus could no longer be improved. The re-filtered sequence data was then remapped onto the initial consensus genome sequence with bowtie2, using the above settings. The genomecov function from BEDtools v2.17.0 (ref. 63) was used to identify gaps and low coverage regions in the resulting alignment.
We attempted to fill the identified gaps, which likely consisted of diverged or difficult-to-assemble regions. For this, we assembled the re-filtered sequence dataset into de novo contigs with the MEGAHIT pipeline v1.1.4 (ref. 64), using a minimum k-mer length of 21, a maximum k-mer length of 63, and k-mer length increments of six. The MEGAHIT contigs were then mapped onto the initial consensus genome sequence with the blastn tool from the NCBI-BLAST+ toolkit. Contigs that covered the gaps identified by BEDtools were incorporated into the initial consensus genome sequence, unless a blast comparison against the NCBI nucleotide database suggested a closer match to non-Nannochloropsis taxa. We repeated the bowtie2 gap-filling steps iteratively, using the previous consensus sequence as reference, until a gap-free consensus was obtained. The re-filtered sequence data were again mapped, the resulting final assembly was visually inspected, and the consensus was corrected where necessary. This was to ensure the fidelity of the consensus sequence, which incorporated de novo-assembled contigs that could potentially be problematic, due to the fragmented nature and deaminated sites of ancient DNA impeding accurate assembly65.
Annotation of the chloroplast genome was carried out with GeSeq v1.77 (ref. 66), using the available annotated Nannochloropsis chloroplast genomes (accession codes provided in Supplementary Table 7). The resulting annotated chloroplast was visualized with OGDRAW v1.3.1 (ref. 67).
The same assembly and annotation methods outlined above were used to reconstruct the mitochondrial palaeogenome sequence, where the initial mapping assembly was based on the N. limnetica mitochondrial sequence (NCBI GenBank accession: NC_022256.1). The final annotation was carried out by comparison against all available annotated Nannochloropsis mitochondrial genomes (accession codes provided in Supplementary Table 7).
If the Nannochloropsis sequences derived from more than one taxon, then alignment to the N. limnetica chloroplast genome could introduce reference bias, which would underestimate the diversity of the Nannochloropsis sequences present. We therefore reconstructed Nannochloropsis chloroplast genomes, but using the six available Nannochloropsis chloroplast genome sequences, including N. limnetica, as seed genomes (accession codes for the reference genomes are provided in Supplementary Table 3). The assembly of the consensus sequences followed the same method outlined above, but with two modifications to account for the mapping rate being too low for complete genome reconstruction based on alignment to the non-N. limnetica reference sequences. First, consensus sequences were called with SAMtools, which does not incorporate reference bases into the consensus at uncovered sites. Second, neither additional gap filling nor manual curation was implemented.
Analysis of ancient DNA damage patterns
We checked for the presence of characteristic ancient DNA damage patterns for sequences aligned to three nuclear genomes: human, Nannochloropsis limnetica and Mycobacterium avium. We further analysed damage patterns for sequences aligned to both the reconstructed N. limnetica composite organellar genomes. Damage analysis was conducted with mapDamage v2.0.8 (ref. 68) using the following settings: –merge-reference-sequences and –length = 160.
Assembly of high- and low-frequency variant consensus sequences
The within-sample variants in each reconstructed organellar palaeogenome was explored by creating two consensus sequences, which included either high- or low-frequency variants at multiallelic sites. For each sample, the initial filtered sequence data were mapped onto the reconstructed Nannochloropsis chloroplast palaeogenome sequence with bowtie2 using default settings. Unmapped and duplicate sequences were removed with SAMtools, as above. We used the BCFtools mpileup, call, and normalize functions to identify the variant sites in the mapped dataset, using the –skip-indels, –variants-only, and –multiallelic-caller options. The resulting alleles were divided into two sets, based on either high- or low-frequency variants. High-frequency variants were defined as those present in the reconstructed reference genome sequence. Both sets were further filtered to only include sites with a quality score of 30 or higher and a coverage of at least half the average coverage of the mapping assembly (minimum coverage: Andøya_LGM_A = 22×, Andøya_LGM_B = 14×). We then generated the high- and low-frequency variant consensus sequences using the consensus function in BCFTools. The above method was repeated for the reconstructed Nannochloropsis mitochondrial genome sequence in order to generate comparable consensus sequences of high- and low-frequency variants (minimum coverage: Andøya_LGM_A = 16×, Andøya_LGM_B = 10×).
Phylogenetic analysis of the reconstructed organellar palaeogenomes
We determined the phylogenetic placement of our high- and low-frequency variant organellar palaeogenomes within Nannochloropsis, using either full mitochondrial and chloroplast genome sequences or three short loci (18S, ITS, rbcL). We reconstructed the 18S and ITS1-5.8S-ITS2 complex using DQ977726.1 (full length) and EU165325.1 (positions 147:1006, corresponding to the ITS complex) as seed sequences following the same approach that was used for the organellar palaeogenome reconstructions, except that the first and last 10 bp were trimmed to account for the lower coverage due to sequence tiling. We then called high and low variant consensus sequences as described above.
We created six alignments using available sequence data from NCBI GenBank (Supplementary Data 4) with the addition of: (1 + 2) the high- and low-frequency variant chloroplast or mitochondrial genome consensus sequences, (3) an ~1100 bp subset of the chloroplast genome for the rbcL alignment, (4 + 5) ~1800 and ~860 bp subsets of the nuclear multicopy complex for the 18S and ITS alignments, respectively, and (6) the reconstructed chloroplast genome consensus sequences derived from the alternative Nannochloropsis genome starting points. Full details on the coordinates of the subsets are provided in Supplementary Data 4. We generated alignments using MAFFT v7.427 (ref. 69) with the maxiterate = 1000 setting, which was used for the construction of a maximum likelihood tree in RAxML v8.1.12 (ref. 70) using the GTRGAMMA model and without outgroup specified. We assessed branch support using 1000 replicates of rapid bootstrapping.
Nannochloropsis variant proportions and haplogroup diversity estimation
To estimate major haplogroup diversity, we calculated the proportions of high and low variants in the sequences aligned to our reconstructed Nannochloropsis mitochondrial and chloroplast genomes. For each sample, we first mapped the initial filtered sequence data onto the high- and low-frequency variant consensus sequences with bowtie2. To avoid potential reference biases, and for each organellar genome, the sequence data were mapped separately against both frequency consensus sequences. The resulting bam files were then merged with SAMtools merge. We removed exact sequence duplicates, which may have been mapped to different coordinates, from the merged bam file by randomly retaining one copy. This step was replicated five times to examine its impact on the estimated variant proportions. After filtering, remaining duplicate sequences—those with identical mapping coordinates—were removed with SAMtools rmdup. We then called variable sites from the duplicate-removed bam files using BCFTools under the same settings as used in the assembly of the high- and low-frequency variant consensus sequences. We restricted our analyses to transversion-only variable positions to remove the impact of ancient DNA deamination artifacts. For each variable site, the proportion of reference and alternative alleles was calculated, based on comparison to the composite N. limnetica reconstructed organellar palaeogenomes. We removed rare alleles occurring at a proportion of More

Viruses and mammalian data
Viral genomic data
Complete sequences of coronaviruses were downloaded from Genbank44. Sequences labelled with the terms: ‘vaccine’, ‘construct’, ‘vector’, ‘recombinant’ were removed from the analyses. In addition, we removed those associated with experimental infections where possible. This resulted in a total of 3264 sequences for 411 coronavirus species or strains (i.e., viruses below species level on NCBI taxonomy tree). Of those, 88 were sequences of coronavirus species, and 307 sequences of strains (in 25 coronavirus species, with total number of species included = 92). Of our included species, six in total were unclassified Coronavirinae (unclassified coronaviruses).
Selection of potential mammalian hosts of coronaviruses
We processed meta-data accompanying all sequences (including partial sequences but excluding vaccination and experimental infections) of coronaviruses uploaded to GenBank to extract information on hosts (to species level) of these coronaviruses. We supplemented these data with species-level hosts of coronaviruses extracted from scientific publications via the ENHanCEd Infectious Diseases Database (EID2)45. This resulted in identification of 313 known terrestrial mammalian hosts of coronaviruses (regardless of whether a complete genome was available or not, n = 185 mammalian species for which an association with a coronavirus with complete genome was identified). We expanded this set of potential hosts by including terrestrial mammalian species in genera containing at least one known host of coronavirus, and which are known to host one or more other virus species (excluding coronaviruses, information of whether the host is associated with a virus were obtained from EID2). This results in total of 876 mammalian species which were selected.
Quantification of viral similarities
We computed three types of similarities between each two viral genomes as summarised below.
Biases and codon usage
We calculated proportion of each nucleotide of the total coding sequence length. We computed dinucleotide and codon biases46 and codon-pair bias, measured as the codon-pair score46,47 in each of the above sequences. This enabled us to produce for each genome sequence (n = 3264) the following feature vectors: nucleotide bias, dinucleotide bias, codon biased and codon-pair bias.
Secondary structure
Following alignment of sequences (using AlignSeqs function in R package Decipher48), we predicted the secondary structure for each sequence using PredictHEC function in the R package Decipher48. We obtained both states (final prediction), and probability of secondary structures for each sequence. We then computed for each 1% of the genome length both the coverage (number of times a structure was predicted) and mean probability of the structure (in the per cent of the genome considered). This enabled us to generate six vectors (length = 100) for each genome representing: mean probability and coverage for each of three possible structures—Helix (H), Beta-Sheet (E) or Coil (C).
Genome dissimilarity (distance)
We calculated pairwise dissimilarity (in effect a hamming distance) between each two sequences in our set using the function DistanceMatrix in the R package Decipher48. We set this function to penalise gap-to-gap and gap-to-letter mismatches.
Similarity quantification
We transformed the feature (traits) vectors described above into similarities matrices between coronaviruses (species or strains). This was achieved by computing cosine similarity between these vectors in each category (e.g., codon-pair usage, H coverage, E probability). Formally, for each genomic feature (n = 10) presented by vector as described above, this similarity was calculated as follows:

$${mathrm{sim}}_{{mathrm{genomic}}_{mathrm{l}}}left( {s_m,s_n} right) = {mathrm{sim}}_{{mathrm{genomic}}_{mathrm{l}}}left( {{mathbf{V}}_m^{f_l},{mathbf{V}}_n^{f_l}} right) = frac{{mathop {sum}nolimits_{i = 1}^d {left( {{mathbf{V}}_m^{f_l}[i] times {mathbf{V}}_n^{f_l}[i]} right)} }}{{sqrt {mathop {sum}nolimits_{i = 1}^d {{mathbf{V}}_m^{f_l}[i]^2} } times sqrt {mathop {sum}nolimits_{i = 1}^d {{mathbf{V}}_n^{f_l}[i]^2} } }}$$
(1)

where sm and sn are two sequences presented by two feature vectors ({mathbf{V}}_m^{f_l}) and ({mathbf{V}}_n^{f_l}) from the genomic feature space fl (e.g., codon-pair bias) of the dimension d (e.g., d = 100 for H coverage).
We then calculated similarity between each pair of virus strains or species (in each category) as the mean of similarities between genomic sequences of the two virus strains or species (e.g., mean nucleotide bias similarity between all sequences of SARS-CoV-2 and all sequences of MERS-CoV presented the final nucleotide bias similarity between SARS-CoV-2 and MERS-CoV). This enabled us to generate 11 genomic features similarity matrices (the above 10 features represented by vectors and genomic similarity matrix) between our input coronaviruses. Supplementary Fig. 1 illustrates the process.
Similarity network fusion (SNF)
We applied SNF49 to integrate the following similarities in order to reduce our viral genomic feature space: (1) nucleotide, dinucleotide, codon and codon-pair usage biases were combined into one similarity matrix—genome bias similarity. And (2) Helix (H), Beta-Sheet (E) or Coil (C) mean probability and coverage similarities (six in total) were combined into one similarity matrix—secondary structure similarity.
SNF applies an iterative nonlinear method that updates every similarity matrix according to the other matrices via nearest neighbour approach and is scalable and is robust to noise and data heterogeneity. The integrated matrix captures both shared and complementary information from multiple similarities.
Quantification of mammalian similarities
We calculated a comprehensive set of mammalian similarities. Table 3 summarises these similarities and provides justification for inclusion. Supplementary Note 1 provides full details.
Table 3 Mammalian phylogenetic, ecological and geospatial similarities.
Full size table

Quantification of network similarities
Network construction
We processed meta-data accompanying all sequences (including partial genome but excluding vaccination and experimental infections) of coronaviruses uploaded to Genbank44 (accessed 4 May 2020) to extract information on hosts (to species level) of these coronaviruses. We supplemented these data with virus–host associations extracted from publications via the EID2 Database45. This resulted in 1669 associations between 1108 coronaviruses and 545 hosts (including non-mammalian hosts). We transformed these associations into a bipartite network linking species and strains of coronaviruses with their hosts.
Quantification of topological features
The above constructed network summarises our knowledge to date of associations between coronaviruses and their hosts, and its topology expresses patterns of sharing these viruses between various hosts and host groups. Our analytical pipeline captures this topology, and relations between nodes in our network, by means of node embeddings. This approach encodes each node (here either a coronavirus or a host) with its own vector representation in a continuous vector space, which, in turn, enables us to calculate similarities between two nodes based on this representation.
We adopted DeepWalk23 to compute vectorised representations for our coronaviruses and hosts from the network connecting them. DeepWalk23 uses truncated random walks to get latent topological information of the network and obtains the vector representation of its nodes (in our case coronaviruses and their hosts) by maximising the probability of reaching a next node (i.e., probability of a virus–host association) given the previous nodes in these walks (Supplementary Note 2 lists further details).
Similarity calculations
Following the application of DeepWalk to compute the latent topological representation of our nodes, we calculated the similarity between two nodes in our network—n (vectorised as N) and m (vectorised as M), by using cosine similarity as follows24,25:

$${mathrm{sim}}_{{mathrm{network}}}left( {n,m} right) = {mathrm{sim}}_{{mathrm{network}}}left( {{mathbf{M}},{mathbf{N}}} right) = frac{{mathop {sum}nolimits_{i = 1}^d {left( {m_i times n_i} right)} }}{{sqrt {mathop {sum}nolimits_{i = 1}^d {m_i^2} } times sqrt {mathop {sum}nolimits_{i = 1}^d {n_i^2} } }}$$
(2)

where d is the dimension of the vectorised representation of our nodes: M, N; and mi and ni are the components of vectors M and N, respectively.
Similarity learning meta-ensemble—a multi-perspective approach
Our analytical pipeline stacks 12 similarity learners into testable meta-ensembles. The constituent learners can be categorised by the following three ‘points of view’ (see also Supplementary Fig. 4 for a visual description):
Coronaviruses—the virus point of view
We assembled three models derived from (a) genome similarity, (b) genome biases and (c) genome secondary structure. Each of these learners gave each coronavirus–mammalian association (( {v_i to m_j} )) a score, termed confidence, based on how similar the coronavirus vi is to known coronaviruses of mammalian species mj, compared to how similar vi is to all included coronaviruses. In other words, if vi is more similar (e.g., based on genome secondary structure) to coronaviruses observed in host mj than it is similar to all coronaviruses (both observed in mj and not), then the association (v_i to m_j) is given a higher confidence score. Conversely, if vi is somewhat similar to coronaviruses observed in mj, and also somewhat similar to viruses not known to infect this particular mammal, then the association (v_i to m_j) is given a medium confidence score. The association (v_i to m_j) is given a lower confidence score if vi is more similar to coronaviruses not known to infect mj than it is similar to coronaviruses observed in this host.
Formally, given an adjacency matrix A of dimensions (left| {mathbf{V}} right| times left| {mathbf{M}} right|) where (left| {mathbf{V}} right|) is number of coronaviruses included in this study (for which a complete genome could be found), and (left| {mathbf{M}} right|) is number of included mammals, such that for each (v_i in {mathbf{V}}) and (m_j in {mathbf{M}}), aij = 1 if an association is known to exist between the virus and the mammal, and 0 otherwise. Then for a similarity matrix simviral corresponding to each of the similarity matrices calculated above, a learner from the viral point of view is defined as follows24,25:

$${mathrm{confidence}}_{{mathrm{viral}}}( {v_i to m_j} ) = frac{{mathop {sum}nolimits_{l = 1,,l ne i}^{left| {mathbf{V}} right|} {( {{mathrm{sim}}_{{mathrm{viral}}}( {v_i,,v_l} ) times a_{lj}} )} }}{{mathop {sum}nolimits_{l = 1,,l ne i}^{left| {mathbf{V}} right|} {{mathrm{sim}}_{{mathrm{viral}}}left( {v_i,,v_l} right)} }}$$
(3)

Mammals—the host point of view
We constructed seven learners from the similarities summarised in Table 3. Each of these learners calculated for every coronavirus–mammalian association (( {v_i to m_j})) a confidence score based on how similar the mammalian species mj is to known hosts of the coronavirus vi, compared to how similar mj is to mammals not associated with vi. For instance, if mj is phylogenetically close to known hosts of vi, and also phylogenetically distant to mammalian species not known to be associated with this coronavirus, then the phylogenetic similarly learner will assign (v_i to m_j) a higher confidence score. However, if mj does not overlap geographically with known hosts of vi, then the geographical overlap learner will assign it a low (in effect 0) confidence score.
Formally, given the above-defined adjacency matrix A, and a similarity matrix simmammalian corresponding to each of the similarity matrices summarised in Table 3, a learner from the mammalian point of view is defined as follows24,25:

$${mathrm{confidence}}_{{mathrm{mammalian}}}( {v_i to m_j} ) = frac{{mathop {sum}nolimits_{l = 1,,l ne j}^{left| {mathbf{M}} right|} {( {{mathrm{sim}}_{{mathrm{mammalian}}}( {m_j,,m_l} ) times a_{il}} )} }}{{mathop {sum}nolimits_{l = 1,,l ne j}^{left| {mathbf{M}} right|} {{mathrm{sim}}_{{mathrm{mammalian}}}( {m_j,,m_l} )} }}$$
(4)

Network—the network point of view
We integrated two learners based on network similarities—one for mammals and one for coronaviruses. Formally, given the adjacency matrix A, our two learners from the network point of view as defined as follows24:

$${mathrm{confidence}}_{{mathrm{network}}_{mathbf{V}}}( {v_i to m_j} ) = frac{{mathop {sum}nolimits_{l = 1,,l ne i}^{left| {mathbf{V}} right|} {( {{mathrm{sim}}_{{mathrm{network}}}( {v_i,,v_l} ) times a_{lj}} )} }}{{mathop {sum}nolimits_{l = 1,,l ne i}^{left| {mathbf{V}} right|} {{mathrm{sim}}_{{mathrm{network}}}( {v_i,,v_l} )} }};;$$
(5)

$${mathrm{confidence}}_{{mathrm{network}}_{mathbf{M}}}( {v_i to m_j} ) = frac{{mathop {sum}nolimits_{l = 1,,l ne j}^{left| {mathbf{M}} right|} {( {{mathrm{sim}}_{{mathrm{network}}}( {m_j,,m_l} ) times a_{il}} )} }}{{mathop {sum}nolimits_{l = 1,,l ne j}^{left| {mathbf{M}} right|} {{mathrm{sim}}_{{mathrm{network}}}( {m_j,,m_l} )} }}$$
(6)

Ensemble construction
We combined the learners described above by stacking them into ensembles (meta-ensembles) using Stochastic Gradient Boosting (GBM). The purpose of this combination is to incorporate the three points of views, as well as varied aspects of the coronaviruses and their mammalian potential hosts, into a generalisable, robust model50. We selected GBM as our stacking algorithm following an assessment of seven machine-learning algorithms using held-out test sets (20% of known associations randomly selected, N = 5—Supplementary Fig. 14). In addition, GBM is known for its ability to handle non-linearity and high-order interactions between constituent learners51, and have been used to predict reservoirs of viruses46 and zoonotic hot-spots51.We performed the training and optimisation (tuning) of these ensembles using the caret R Package52.
Sampling
Our GBM ensembles comprised 100 replicate models. Each model was trained with balanced random samples using tenfold cross-validation (Supplementary Fig. 4). Final ensembles were generated by taking mean predictions (probability) of constituent models. Predictions were calculated form the mean probability at three cut-offs: >0.5 (standard), >0.75 and ≥0.9821. SD from mean probability was also generated and its values subtracted/added to predictions, to illustrate variation in the underlying replicate models.
Validation and performance estimation
We validated the performance of our analytical pipeline externally against 20 held-out test sets. Each test set was generated by splitting the set of observed associations between coronaviruses and their hosts into two random sets: a training set comprising 85% of all known associations and a test set comprising 15% of known associations. These test sets were held-out throughout the processes of generating similarity matrices; similarity learning, and assembling our learners, and were only used for the purposes of estimating performance metrics of our analytical pipeline. This resulted in 20 runs in which our ensemble learnt using only 85% of observed associations between our coronaviruses and their mammalian hosts. For each run, we calculated three performance metrics based on the mean probability across each set of 100 replicate models of the GBM meta-ensembles: AUC, true skill statistics (TSS) and F-score.
AUC is a threshold-independent measure of model predictive performance that is commonly used as a validation metric for host–pathogen predictive models21,46. Use of AUC has been criticised for its insensitivity to absolute predicted probability and its inclusion of a priori untenable prediction51,53, and so we also calculated the TSS (TSS = sensitivity + specificity − 1)54. F-score captures the harmonic mean of the precision and recall and is often used with uneven class distribution. Our approach is relaxed with respect to false positives (unobserved associations), hence the low F-score recorded overall.
We selected three probability cut-offs for our meta-ensemble: 0.50, 0.75 and 0.9821. One extreme of our cut-off range (0.5) maximises the ability of our ensemble to detect known associations (higher AUC, lower F-score). The other (0.9821) is calculated so that 90% of known positives are captured by our ensemble, while reducing the number of additional associations predicted (higher F-score, lower AUC).
Changes in network structure
We quantified the diversity of the mammalian hosts of each coronavirus in our input by computing mean phylogenetic distance between these hosts. Similarly, we captured the diversity of coronaviruses associated with each mammalian species by calculating mean (hamming) distance between the genomes of these coronaviruses. We termed these two metrics: mammalian diversity per virus and viral diversity per mammal, respectively. We aggregated both metrics at the network level by means of simple average. This enabled us to quantify changes in these diversity metrics, at the level of network, with addition of predicted links at three probability cut-offs: >0.5, >0.75 and ≥0.9821.
In addition, we captured changes in the structure of the bipartite network linking CoVs with their mammalian hosts, with the addition of predicted associations, by computing a comprehensive set of structural properties (Supplementary Note 3) at the probability cut-offs mentioned above, and comparing the results with our original network. Here we ignore properties that deterministically change with the addition of links (e.g., degree centrality, connectance; Supplementary Table 2 lists all computed metrics and changes in their values). Instead, we focus on non-trivial structural properties. Specifically, we capture changes in network stability, by measuring its nestedness55,56,57; and we quantify non-independence in interaction patterns by means of C-score58. Supplementary Note 3 provides full definition of these concepts as well as other metrics we computed for our networks.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article. More

1.
Ripple, W. J. et al. Status and ecological effects of the world’s largest carnivores. Science 343, 1241484 (2014).
2.
Macdonald, E. A. et al. Conservation inequality and the charismatic cat: Felis felicis. Glob. Ecol. Conserv. 3, 851–866 (2015).
Article  Google Scholar 

3.
Carter, N. H. & Linnell, J. D. C. Co-adaptation is key to coexisting with large carnivores. Trends. Ecol. Evol. 31, 575–578 (2016).
PubMed  Article  Google Scholar 

4.
Inskip, C. & Zimmermann, A. Human-felid conflict: a review of patterns and priorities worldwide. Oryx 43, 18–34 (2009).
Article  Google Scholar 

5.
Macdonald, D. W., Mosser, A. & Gittleman, J. L. Felid society. In Biology and Conservation of Wild Felids (eds Macdonald, D. W. & Loveridge, A. J.) 125–160 (Oxford University Press, Oxford, 2010).
Google Scholar 

6.
Treves, A. & Karanth, K. U. Human-carnivore conflict and perspectives on carnivore management worldwide. Conserv. Biol. 17, 1491–1499 (2003).
Article  Google Scholar 

7.
Lute, M. L., Carter, N. H., López-Bao, J. V. & Linnell, J. D. C. Conservation professionals agree on challenges to coexisting with large carnivores but not on solutions. Biol. Conserv. 218, 223–232 (2018).
Article  Google Scholar 

8.
Chapron, G. et al. Recovery of large carnivores in Europe’s modern human-dominated landscapes. Science 346, 1517–1519 (2014).
ADS  CAS  PubMed  Article  Google Scholar 

9.
Van der Meer, E., Fritz, H., Blinston, P. & Rasmussen, G. S. A. Ecological trap in the buffer zone of a protected area: effects of indirect anthropogenic mortality on the African wild dog (Lycaon pictus). Oryx 48, 285–293 (2014).
Article  Google Scholar 

10.
Khorozyan, I., Ghoddousi, A., Soofi, M. & Waltert, M. Big cats kill more livestock when wild prey reaches a minimum threshold. Biol. Conserv. 192, 268–275 (2015).
Article  Google Scholar 

11.
Cozzi, G. et al. Anthropogenic food resources foster the coexistence of distinct life history strategies: year-round sedentary and migratory brown bears. J. Zool. 300, 142–150 (2016).
Article  Google Scholar 

12.
Schuette, P., Wagner, A. P., Wagner, M. E. & Creel, S. Occupancy patterns and niche partitioning within a diverse carnivore community exposed to anthropogenic pressures. Biol. Conserv. 158, 301–312 (2013).
Article  Google Scholar 

13.
Oriol-Cotterill, A., Macdonald, D. W., Valeix, M., Ekwanga, S. & Frank, L. G. Spatiotemporal patterns of lion space use in a human-dominated landscape. Anim. Behav. 101, 27–39 (2015).
Article  Google Scholar 

14.
Suraci, J. P. et al. Behavior-specific habitat selection by African lions may promote their persistence in a human-dominated landscape. Ecology 100, 1–11 (2019).
Article  Google Scholar 

15.
Kuijper, D. P. J. et al. Paws without claws? Ecological effects of large carnivores in anthropogenic landscapes. Proc. R. Soc. B Biol. Sci. 283, 20161625 (2016).
Article  Google Scholar 

16.
Carter, N. H., Shrestha, B. K., Karki, J. B., Pradhan, N. M. B. & Liu, J. Coexistence between wildlife and humans at fine spatial scales. Proc. Natl. Acad. Sci. 109, 15360–15365 (2012).
ADS  CAS  PubMed  Article  Google Scholar 

17.
Odden, M., Athreya, V., Rattan, S. & Linnell, J. D. C. Adaptable neighbours: Movement patterns of GPS-collared leopards in human dominated landscapes in India. PLoS ONE 9, .e112044. https://doi.org/10.1002/ecs2.1408 (2014).

18.
Wang, Y., Allen, M. L. & Wilmers, C. C. Mesopredator spatial and temporal responses to large predators and human development in the Santa Cruz Mountains of California. Biol. Conserv. 190, 23–33 (2015).
Article  Google Scholar 

19.
Wheat, R. E. & Wilmers, C. C. Habituation reverses fear-based ecological effects in brown bears (Ursus arctos). Ecosphere 7, (2016).

20.
Gaynor, K. M., Hojnowski, C. E., Carter, N. H. & Brashares, J. S. The influence of human disturbance on wildlife nocturnality. Science 360, 1232–1235 (2018).
ADS  CAS  PubMed  Article  Google Scholar 

21.
Valeix, M. et al. How key habitat features influence large terrestrial carnivore movements: waterholes and African lions in a semi-arid savanna of north-western Zimbabwe. Landsc. Ecol. 25, 337–351 (2010).
Article  Google Scholar 

22.
Dickson, B. G., Jenness, J. S. & Beier, P. Influence of vegetation, topography, and roads on cougar movement in southern California. J. Wildl. Manag. 69, 264–276 (2005).
Article  Google Scholar 

23.
Elliot, N. B., Cushman, S. A., Macdonald, D. W. & Loveridge, A. J. The devil is in the dispersers: predictions of landscape connectivity change with demography. J. Appl. Ecol. 51, 1169–1178 (2014).
Article  Google Scholar 

24.
Valeix, M., Hemson, G., Loveridge, A. J., Mills, G. & Macdonald, D. W. Behavioural adjustments of a large carnivore to access secondary prey in a human-dominated landscape. J. Appl. Ecol. 49, 73–81 (2012).
Article  Google Scholar 

25.
Farhadinia, M. S. et al. Vertical relief facilitates spatial segregation of a high density large carnivore population. Oikos 129, 346–355 (2020).
Article  Google Scholar 

26.
Grant, J., Hopcraft, C., Sinclair, A. R. E. & Packer, C. Planning for success: Serengeti lions seek prey accessibility rather than abundance. J. Anim. Ecol. 74, 559–566 (2005).
Article  Google Scholar 

27.
Laundré, J. W. & Hernández, L. Total energy budget and prey requirements of free-ranging coyotes in the Great Basin desert of the western United States. J. Arid. Environ. 55, 675–689 (2003).
ADS  Article  Google Scholar 

28.
Miller, J. R. B., Jhala, Y. V., Jena, J. & Schmitz, O. J. Landscape-scale accessibility of livestock to tigers: implications of spatial grain for modeling predation risk to mitigate human-carnivore conflict. Ecol. Evol. 5, 1354–1367 (2015).
PubMed  PubMed Central  Article  Google Scholar 

29.
Blake, L. W. & Gese, E. M. Resource selection by cougars: influence of behavioral state and season. J. Wildl. Manag. 80, 1205–1217 (2016).
Article  Google Scholar 

30.
Broekhuis, F., Grünewälder, S., McNutt, J. W. & Macdonald, D. W. Optimal hunting conditions drive circalunar behavior of a diurnal carnivore. Behav. Ecol. 25, 1268–1275 (2014).
Article  Google Scholar 

31.
Klaassen, B. & Broekhuis, F. Living on the edge: multiscale habitat selection by cheetahs in a human-wildlife landscape. Ecol. Evol. 8, 7611–7623 (2018).
PubMed  PubMed Central  Article  Google Scholar 

32.
Gese, E. M. et al. Cross-fostering as a conservation tool to augment endangered carnivore populations. J. Mammal. 99, 1033–1041 (2018).
Article  Google Scholar 

33.
Naha, D. et al. Ranging, activity and habitat use by tigers in the mangrove forests of the Sundarban. PLoS ONE 11, e0152119 (2016).
PubMed  PubMed Central  Article  CAS  Google Scholar 

34.
Karanth, K. K., Nichols, J. D., Hines, J. E., Karanth, K. U. & Christensen, N. L. Patterns and determinants of mammal species occurrence in India. J. Appl. Ecol. https://doi.org/10.1111/j.1365-2664.2009.01710.x (2009).
Article  Google Scholar 

35.
Athreya, V. et al. Spotted in the news: using media reports to examine leopard distribution, depredation, and management practices outside protected areas in Southern India. PLoS ONE 10, e0142647 (2015).
PubMed  PubMed Central  Article  CAS  Google Scholar 

36.
Naha, D., Sathyakumar, S. & Rawat, G. S. Understanding drivers of human-leopard conflicts in the Indian Himalayan region: spatio-temporal patterns of conflicts and perception of local communities towards conserving large carnivores. PLoS ONE 13, 1–19 (2018).
Article  CAS  Google Scholar 

37.
Jacobson, A. P. et al. Leopard (Panthera pardus) status, distribution, and the research efforts across its range. PeerJ, 4, 1–28 (2016).

38.
Naha, D. et al. Landscape predictors of human-leopard conflicts within multi-use areas of the Himalayan Region. Sci. Rep. 10, 11129. https://doi.org/10.1038/s41598-020-67980-w (2020).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

39.
Odden, M. & Wegge, P. Spacing and activity patterns of leopards (Panthera pardus) in the Royal Bardia National Park, Nepal. Wildl. Biol. 11, 145–152 (2005).
Article  Google Scholar 

40.
Farhadinia, M. S., Johnson, P. J., Macdonald, D. W. & Hunter, L. T. B. Anchoring and adjusting amidst humans: ranging behavior of Persian leopards along the Iran-Turkmenistan borderland. PLoS ONE 13, e0196602 (2018).
PubMed  PubMed Central  Article  CAS  Google Scholar 

41.
Miquelle, D. G. et al. Tigers and wolves in the Russian Far East: competitive exclusion, functional redundancy and conservation implications. In Large Carnivores and the Conservation of Biodiversity (eds Ray, J. C. et al.) 179–207 (Island Press, Washington, 2005).
Google Scholar 

42.
Powell, R. A. Animal home ranges and territories and home range estimators. In Boitani (ed. L. & Fuller, T.K. ) 65–110 (Research Techniques in Anim. Ecol. Controversies and Consequences, Columbia University Press, New York, 2000).
Google Scholar 

43.
Mizutani, F. & Jewell, P. A. Home-range and movements of leopards (Panthera pardus) on a livestock ranch in Kenya. J. Zool. 244, 269–286 (1998).
Article  Google Scholar 

44.
Ray-Brambach, R. R., Stommel, C. & Rödder, D. Home ranges, activity patterns and habitat preferences of leopards in Luambe National Park and adjacent Game Management Area in the Luangwa Valley, Zambia. Mamm. Biol. 92, 102–110 (2018).
Article  Google Scholar 

45.
Karelus, D. L. et al. Effects of environmental factors and landscape features on movement patterns of Florida black bears. J. Mammal. 98, 1463 (2017).
Article  Google Scholar 

46.
Athreya, V., Odden, M., Linnell, J. D. C., Krishnaswamy, J. & Karanth, K. U. A cat among the dogs: Leopard (Panthera pardus) diet in a human-dominated landscape in western Maharashtra, India. Oryx 50, 156–162 (2016).
Article  Google Scholar 

47.
Kshettry, A., Vaidyanathan, S. & Athreya, V. Diet selection of leopards (Panthera pardus) in a human-use landscape in North-Eastern India. Trop. Conserv. Sci. 11, 194008291876463 (2018).
Article  Google Scholar 

48.
Ciuti, S. et al. Effects of humans on behaviour of wildlife exceed those of natural predators in a landscape of fear. PLoS ONE 7(11), e50611. https://doi.org/10.1371/journal.pone.0050611 (2012).

49.
Tucker, M. A. et al. Moving in the anthropocene: global reductions in terrestrial mammalian movements. Science 359, 466 (2018).
ADS  CAS  PubMed  Article  Google Scholar 

50.
Broekhuis, F., Madsen, E. K. & Klaassen, B. Predators and pastoralists: how anthropogenic behavior inside wildlife areas influence carnivore space use and movement behavior. Anim. Conserv. https://doi.org/10.1111/acv.12483 (2019).
Article  Google Scholar 

51.
Filla, M. et al. Habitat selection by Eurasian lynx (Lynx lynx) is primarily driven by avoidance of human activity during day and prey availability during night. Ecol. Evol. 7, 6367–6381 (2017).
PubMed  PubMed Central  Article  Google Scholar 

52.
Du Preez, B., Hart, T., Loveridge, A. J. & Macdonald, D. W. Impact of risk on animal behaviour and habitat transition probabilities. Anim. Behav. 100, 22–37 (2015).
Article  Google Scholar 

53.
Van Cleave, E. K. et al. Diel patterns of movement activity and habitat use by leopards (Panthera pardus pardus) living in a human-dominated landscape in central Kenya. Biol. Conserv. 226, 224–237 (2018).
Article  Google Scholar 

54.
Durant, S. M. Living with the enemy: avoidance of hyenas and lions by cheetahs in the Serengeti. Behav. Ecol. 11, 624–632 (2000).
Article  Google Scholar 

55.
Broekhuis, F., Cozzi, G., Valeix, M., McNutt, J. W. & Macdonald, D. W. Risk avoidance in sympatric large carnivores: reactive or predictive?. J. Anim. Ecol. 82, 1098–1105 (2013).
PubMed  Article  Google Scholar 

56.
Packer, C., Swanson, A., Ikanda, D. & Kushnir, H. Fear of darkness, the full moon and the nocturnal ecology of African lions. PLoS ONE 6(7), e22285. https://doi.org/10.1371/journal.pone.0022285 (2011).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

57.
Schuette, P., Creel, S. & Christianson, D. (2013) Coexistence of African lions, livestock, and people in a landscape with variable human land use and seasonal movements. Biol. Conserv. 157, 148–154 (2013).
Article  Google Scholar 

58.
Acharya, K. P., Paudel, P. K., Neupane, P. R. & Köhl, M. Human–wildlife conflicts in Nepal: patterns of human fatalities and injuries caused by large mammals. PLoS ONE 11, e0161717 (2016).
PubMed  PubMed Central  Article  CAS  Google Scholar 

59.
Dhanwatey, H. S. et al. Large carnivore attacks on humans in Central India:aA case study from the Tadoba-Andheri Tiger Reserve. Oryx 47(2), 221–227. https://doi.org/10.1017/S0030605311001803 (2013).
Article  Google Scholar 

60.
Ngoprasert, D., Lynam, A. J. & Gale, G. A. Human disturbance affects habitat use and behaviour of Asiatic leopard (Panthera pardus) in Kaeng Krachan National Park, Thailand. Oryx 41, 343–351 (2007).
Article  Google Scholar 

61.
Carter, N., Jasny, M., Gurung, B. & Liu, J. Impacts of people and tigers on leopard spatiotemporal activity patterns in a global biodiversity hotspot. Glob. Ecol. Conserv. 3, 149–162 (2015).
Article  Google Scholar 

62.
Bhattacharjee, A. & Parthasarathy, N. Coexisting with large carnivores: a case study from Western Duars, India. Hum. Dimens. Wildl. 18, 20–31 (2013).
Article  Google Scholar 

63.
Naha, D., Sathyakumar, S., Dash, S., Chettri, A. & Rawat, G. S. Assessment and prediction of spatial patterns of human-elephant conflicts in changing land cover scenarios of a human-dominated landscape in North Bengal. PLoS ONE 14, 1–19 (2019).
Article  CAS  Google Scholar 

64.
Kreeger, T. K. Handbook of Wildlife Chemical immobilization. International Wildlife Vet. Services Inc. Post Box 37, Larammie, WY, USA, (1996).

65.
Calenge, C. The package “adehabitat” for the R software: a tool for the analysis of space and habitat use by animals. Ecol. Modell. 197, 516–519 (2006).
Article  Google Scholar 

66.
Kie, J. G., Terry-Bowyer, R., Nicholson, M. C., Boroski, B. B. & Loft, E. R. Landscape heterogeneity at differing scales: Effects on spatial distribution of mule deer. Ecology 83, 530–544 (2002).
Article  Google Scholar 

67.
Kie, J. G. et al. The home-range concept: are traditional estimators still relevant with modern telemetry technology?. Philos. Trans. R. Soc. B. Biol. Sci. 365, 2221–2231 (2010).
Article  Google Scholar 

68.
Michelot, T., Langrock, R. & Patterson, T. A. moveHMM: an R package for the statistical modelling of animal movement data using hidden Markov models. Methods Ecol Evol 7, 1308–1315 (2016).
Article  Google Scholar 

69.
R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, 2019).
Google Scholar 

70.
Bates, D., Kliegl, R., Vasishth, S. & Baayen, H. Parsimonious Mixed Models. ArXiv150604967 Stat. http://arxiv.org/abs/1506.04967 (2015).

71.
Johnson, D. H. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61, 65–71 (1980).
Article  Google Scholar 

72.
Pebesma, E. J. & Bivand, R. S. Classes and methods for spatial data in R. R News 5(2), 9–13 (2005).
Google Scholar 

73.
Esri Inc. ArcGIS Pro (Version 2.4.2). Esri Inc. https://www.esri.com/en-us/arcgis/products/arcgis-pro/ (2019).

74.
Dormann, C. F. et al. Methods to account for spatial autocorrelation in the analysis of species distributional data: a review. Ecography 30, 609–628 (2007).
Article  Google Scholar 

75.
Burnham, K. P. & Anderson, D. R. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach (Springer, Berlin, 2002).
Google Scholar  More

Nonnutritive sweeteners promote conjugative transfer
To test the effect of nonnutritive sweeteners on the conjugative transfer of ARGs, both intra- and intergenus-transfer experiments (model I) were first conducted, in which the bacteria were exposed to various concentrations of four commonly used nonnutritive sweeteners (SAC, SUC, ASP, and ACE-K) for 8 h at room temperature. Notably, in both mating systems, the whole concentration range (from 0.03 to 300 mg/L) of three sweeteners (SUC, ASP, and ACE-K) caused a significant concentration-dependent increase (p = 0.00017 ~ 0.047, Fig. S1a, b); Pearson correlation analysis was shown in Table S3 in conjugative transfer compared to the control (Fig. 1a, b). The intragenus (donor Escherichia coli K-12 LE392 and recipient E. coli K-12 MG1655) spontaneous conjugative transfer frequency was (1.9 ± 0.2) × 10−4 transconjugants per recipient cell (Fig. S2). However, the conjugative transfer frequencies were significantly enhanced by the sweeteners SUC, ASP, and ACE-K at 0.3 mg/L or above. For example, SUC, ASP, and ACE-K at 300 mg/L enhanced the conjugative frequencies by 1.5- (p = 0.00027), 4.1- (p = 0.000000089), and 3.4-fold (p = 0.0000020), respectively (Fig. 1a). In contrast, SAC did not significantly change the conjugative transfer frequency in the conjugation system (p = 0.200 ~ 0.670, Fig. 1a). For intergenus conjugation (donor E. coli K-12 LE392 and recipient Pseudomonas alloputida), all sweeteners at concentrations of 3 mg/L or higher (except for SAC) were seen to promote the conjugative transfer of the donor RP4 plasmid to recipients of different genera (p = 0.000047 ~0.042, Fig. 1b). SUC, ASP, and ACE-K at 300 mg/L caused a great increase in conjugative transfer, by 2.6- (p = 0.0000020), 4.1- (p = 0.000036), and 4.2-fold (p = 0.000019), respectively (Fig. 1b). It should be noted that the enhanced transfer frequencies were associated with the increased number of colonies on selective transconjugant plates, rather than decreased recipient numbers (Fig. S3).
Fig. 1: Nonnutritive sweeteners (SAC, SUC, ASP, and ACE-K) promoted RP4 plasmid-mediated conjugative transfer.

a Fold changes in conjugative ARG transfer within genera. At high concentrations ( >0.3 mg/L), all tested sweeteners (except for SAC) promoted conjugation (N = 6; ANOVA, p  More

We empirically analyzed1 an industrial-scale ‘Dai’ fishery (2000/2001–2014/2015) presenting the signatures of indiscriminate fishing effects on the Tonle Sap’s fish community. Halls and Hortle2 suggest that apparent recent changes in Tonle Sap’s fish catch are more likely to reflect changing hydrological conditions than fishing-down effects, possibly caused by climate change and hydropower development. In addition, they question (1) the use of the Dai fishery data from 2000/01 onwards, as the fishery has been assessed since 19943,4; (2) the Dai data being generated from a ‘standardized biological catch assessment’; (3) the explanation of compensatory response from small-bodied species in stabilizing the Dai seasonal catches; and (4) the mean fish weight used in1 being subject to ‘sampling related bias’. Finally, they claim that fishing effort from Tonle Sap may have declined as a result of fishing lot removal since 2012, and conclude that our findings may distract attention from irreversible and growing threats to fisheries caused by ongoing hydropower dam developments. We appreciate Halls and Hortle’s contributions and the opportunity to discuss these issues.
To begin with, we re-analyzed our data and quantified temporal trends in species’ catch while accounting for both Flood Index (FI) suggested by Halls and Hortle and water temperature effects. To do this, we computed FI using mean daily water level data from Kampong Loung (Tonle Sap Lake) following4 and mean annual water temperature at Prek Kdam in Tonle Sap River where the Dai fishery operates. We then computed for each species a linear model expressing species abundance as a function of the two drivers, and assessed temporal trends in the residuals of these models (i.e. the part of variation not explained by FI and water temperature). The results indicated that 50% of species still showed a declining trend in residuals, with declines significantly more pronounced for large-bodied species (p-value = 0.016, Fig. 1). Here, we reiterate that our results support the findings of others, stressing the increasing contribution of catches of small-bodied species and the decline in the average fish size of the lower Mekong fisheries, both signs of overfishing5,6,7 and indiscriminate fishing effects8,9.
Figure 1

Temporal trend in residuals from the linear models relating species’ catch to Flood Index and water temperature against (a) log-transformed fish species’ maximum total length and (b) trophic level. The results indicated that 50% of species showed a declining trend in residuals, and this decline was significantly stronger for large-bodied fish species (p-value = 0.016).

Full size image

We analyzed the Dai data using 14 rows (Dai row 2–15) from 2000/01 onwards because of changes in sampling schemes and sampling intensity applied to assess the fishery. From 2000/01 onwards, Dai row 1 was discontinued, and it was only from that time that sampling intensity was relatively stable (see Halls et al. 2013, Table 114) with Dai rows and their location unchanged. Although variation in the use of net types and mesh sizes may happen, the Dai gear is long-lasting of about 7 years on average10 and those gears are used seasonally to exploit fish systematically. This makes it seasonally comparable, particularly when the catch is assessed for the entire fishery.
By re-analysing the trends of seasonal catch of the three most prolific small species of the genus Henichorynchus, Halls and Hortle showed no compelling evidence of a compensatory response by small species, but did not consider the increasing trends of other small-bodied species (e.g. Labiobarbus lineatus) that were reported in the Dai catch (see Ngor et al.’s Supplementary Information Table S61). Moreover, Halls and Hortle claim that fishing effort has declined, referring to the fisheries reform in 201211. This reform has led to the establishment of 516 community fisheries (CFi) countrywide. While such policy changes open larger space for local community participation, Cambodia is suffering from governance challenges to successfully implement this policy. The challenges include unclear roles and responsibilities of stakeholders, poor coordination among resource management agencies, limited decentralization of roles and responsibilities to the CFis, weak capacity of the CFi to manage their local resources, insufficient funding to implement the policy and strong livelihoods dependency of the local communities on fisheries, finally resulting in intensification of fishing effort and conflicts over access rights to fisheries resources12,13,14,15,16,17. Fishers indeed borrow money to buy fishing gears, and felt compelled to catch as many fish as possible to repay their loans and meet the household needs15. There is no data on changes in fishing effort, so the claims made by Halls and Hortle on its decrease since 2012 cannot be empirically confirmed or denied. Our reading of the situation is that the fisheries remain under intense pressure despite attempts to implement improved fisheries governance.
Finally, while Tonle Sap remains one of the world’s largest inland fisheries, the system is facing many challenges, including intensive fishing pressure, with most large-sized Mekong fishes listed as threatened by IUCN and unsustainable fishing identified as a primary threat18. Our results do not imply that other impacts are unimportant, but rather that multiple (and potentially synergistic) stressors likely deteriorate the current status of the Tonle Sap fisheries19,20 and need to be further assessed and considered in the decision-making process. More

1.
Hall-Spencer, J. M. & Moore, P. G. Scallop dredging has profound, long-term impacts on maerl habitats. ICES J. Mar. Sci. 57, 1407–1415 (2000).
Article  Google Scholar 
2.
Eigaard, O. R. et al. The footprint of bottom trawling in European waters: Distribution, intensity, and seabed integrity. ICES J. Mar. Sci. 74, 847–865 (2017).
Article  Google Scholar 

3.
Auster, P. J. et al. The impacts of mobile fishing gear on seafloor habitats in the gulf of maine (Northwest Atlantic): Implications for conservation of fish populations. Rev. Fish. Sci. 4, 185–202 (1996).
ADS  Article  Google Scholar 

4.
Gell, F. R. & Roberts, C. M. Benefits beyond boundaries: The fishery effects of marine reserves. Trends Ecol. Evol. 18, 448–455 (2003).
Article  Google Scholar 

5.
Roberts, C. M. et al. Marine reserves canmitigate and promote adaptation to climate change. Proc. Natl. Acad. Sci. USA 114, 6167–6175 (2017).
ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

6.
Sciberras, M., Jenkins, S. R., Kaiser, M. J., Hawkins, S. J. & Pullin, A. S. Evaluating the biological effectiveness of fully and partially protected marine areas. Environ. Evid. 2, 1–31 (2013).
Article  Google Scholar 

7.
Afonso, P., Schmiing, M., Diogo, H. & Serra, R. With various conservation objectives and targets. ICES J. Mar. Sci. 72, 851–862 (2015).
Article  Google Scholar 

8.
Schmiing, M., Diogo, H., Santos, R. S. & Afonso, P. Marine conservation of multispecies and multi-use areas with various conservation objectives and targets. ICES J. Mar. Sci. 72, 851–862 (2015).
Article  Google Scholar 

9.
Giakoumi, S. et al. Ecological effects of full and partial protection in the crowded Mediterranean Sea: A regional meta-analysis. Sci. Rep. 7, 1–12 (2017).
ADS  CAS  Article  Google Scholar 

10.
Zupan, M. et al. Marine partially protected areas: Drivers of ecological effectiveness. Front. Ecol. Environ. 16, 20 (2018).
ADS  Article  Google Scholar 

11.
Halpern, B. S. The impact of marine reserves: Do reserves work and does reserve size matter. Ecol. Appl. 13, S117–S137 (2003).
Article  Google Scholar 

12.
Pikitch, E. K. et al. Ecosystem-based fishery management. Science (80–) 305, 20 (2004).
Article  Google Scholar 

13.
Claudet, J. et al. Marine reserves: Size and age do matter. Ecol. Lett. 11, 481–489 (2008).
PubMed  Article  PubMed Central  Google Scholar 

14.
Lester, S. E. et al. Biological effects within no-take marine reserves: A global synthesis. Mar. Ecol. Prog. Ser. 384, 33–46 (2009).
ADS  Article  Google Scholar 

15.
Fraschetti, S., Guarnieri, G., Bevilacqua, S., Terlizzi, A. & Boero, F. Protection enhances community and habitat stability: Evidence from a Mediterranean marine protected area. PLoS One 8, 20 (2013).
Google Scholar 

16.
Kerwath, S. E., Winker, H., Götz, A. & Attwood, C. G. Marine protected area improves yield without disadvantaging fishers. Nat. Commun. 4, 1–6 (2013).
Article  Google Scholar 

17.
Edgar, G. J. et al. Global conservation outcomes depend on marine protected areas with five key features. Nature 506, 216–220 (2014).
ADS  CAS  PubMed  Article  PubMed Central  Google Scholar 

18.
Hiddink, J. G. et al. Global analysis of depletion and recovery of seabed biota after bottom trawling disturbance. Proc. Natl. Acad. Sci. USA 114, 8301–8306 (2017).
CAS  PubMed  Article  PubMed Central  Google Scholar 

19.
Lombard, A. T. et al. Key challenges in advancing an ecosystem-based approach to marine spatial planning under economic growth imperatives. Front. Mar. Sci. 6, 20 (2019).
Article  Google Scholar 

20.
Trochta, J. T. et al. Ecosystem-based fisheries management: Perception on definitions, implementations, and aspirations. PLoS One 13, 1–9 (2018).
Google Scholar 

21.
EEA. Marine Protected Areas in Europe’s Seas. An Overview and Perspectives for the Future. (2015). https://doi.org/10.2800/99473.

22.
Mangi, S. C., Rodwell, L. D. & Hattam, C. Assessing the impacts of establishing MPAs on fishermen and fish merchants: The case of Lyme Bay, UK. Ambio 40, 457–468 (2011).
PubMed  PubMed Central  Article  Google Scholar 

23.
Luisetti, T. et al. Coastal and marine ecosystem services valuation for policy and management: Managed realignment case studies in England. Ocean Coast. Manag. 54, 212–224 (2011).
Article  Google Scholar 

24.
Molfese, C., Beare, D. & Hall-Spencer, J. M. Overfishing and the replacement of demersal finfish by shellfish: An example from the english channel. PLoS One 9, 20 (2014).
Google Scholar 

25.
Eno, N. C. et al. Effects of crustacean traps on benthic fauna. ICES J. Mar. Sci. 58, 11–20 (2001).
Article  Google Scholar 

26.
Coleman, R. A., Hoskin, M. G., von Carlshausen, E. & Davis, C. M. Using a no-take zone to assess the impacts of fishing: Sessile epifauna appear insensitive to environmental disturbances from commercial potting. J. Exp. Mar. Bio. Ecol. 440, 100–107 (2013).
Article  Google Scholar 

27.
Lewis, C. F., Slade, S. L., Maxwell, K. E. & Matthews, T. R. Lobster trap impact on coral reefs: Effects of wind-driven trap movement. New Zeal. J. Mar. Freshw. Res. 43, 271–282 (2009).
Article  Google Scholar 

28.
Micheli, F., De Leo, G., Butner, C., Martone, R. G. & Shester, G. A risk-based framework for assessing the cumulative impact of multiple fisheries. Biol. Conserv. 176, 224–235 (2014).
Article  Google Scholar 

29.
Stephenson, F., Mill, A. C., Scott, C. L., Polunin, N. V. C. & Fitzsimmons, C. Experimental potting impacts on common UK reef habitats in areas of high and low fishing pressure. ICES J. Mar. Sci. 74, 1648–1659 (2017).
Article  Google Scholar 

30.
Sinclair, M. & Valdimarsson, G. Responsible fisheries in the marine ecosystem. Fish. Res. 20, 426 (2014).
Google Scholar 

31.
Sheehan, E. V., Stevens, T. F., Gall, S. C., Cousens, S. L. & Attrill, M. J. Recovery of a temperate reef assemblage in a marine protected area following the exclusion of towed demersal fishing. PLoS One 8, 1–12 (2013).
Google Scholar 

32.
Sheehan, E. V. et al. Drawing lines at the sand: Evidence for functional vs visual reef boundaries in temperate Marine Protected Areas. Mar. Pollut. Bull. 76, 194–202 (2013).
CAS  PubMed  Article  PubMed Central  Google Scholar 

33.
Jackson, E. L., Langmead, O., Barnes, M., Tyler-Walters, H. & Hiscock, K. Lyme Bay—A Case Study: Measuring Recovery of Benthic Species, Assessing Potential Spill-Over Effects and Socio-economic Changes. (2008).

34.
Stevens, T. F., Sheehan, E. V., Gall, S. C., Fowell, S. C. & Attrill, M. J. Monitoring benthic biodiversity restoration in Lyme Bay marine protected area: Design, sampling and analysis. Mar. Policy 45, 310–317 (2014).
Article  Google Scholar 

35.
Picton, B. E. & Morrow, C. C. Encyclopedia of Marine Life of Britain and Ireland (The Ulster Museum, Belfast, 2016).
Google Scholar 

36.
Langmead. Lyme Bay—A Case Study: Measuring Recovery of Benthic Species, Assessing Potential Spill-Over Effects and Socio-Economic Changes. 44 (2012).

37.
Bradshaw, C., Collins, P. & Brand, A. R. To what extent does upright sessile epifauna affect benthic biodiversity and community composition?. Mar. Biol. 143, 783–791 (2003).
Article  Google Scholar 

38.
Cocito, S., Ferdeghini, F. & Sgorbini, S. Pentapora fascialis (Pallas) [Cheilostomata: Ascophora] colonization of one sublittoral rocky site after sea-storm in the northwestern mediterranean. Hydrobiologia 375–376, 59–66 (1998).
Article  Google Scholar 

39.
Eggleston, D., Lipcius, R., Miller, D. & Coba-Cetina, L. Shelter scaling regulates survival of juvenile Caribbean spiny lobster Panulirus argus. Mar. Ecol. Prog. Ser. 62, 79–88 (1990).
ADS  Article  Google Scholar 

40.
Pirtle, J. L., Eckert, G. L. & Stoner, A. W. Habitat structure influences the survival and predator-prey interactions of early juvenile red king crab Paralithodes camtschaticus. Mar. Ecol. Prog. Ser. 465, 169–184 (2012).
ADS  Article  Google Scholar 

41.
Gall, S. C. et al. The impact of potting for crustaceans on temperate rocky reef habitats: Implications for management. Mar. Environ. Res. 162, 105134 (2020).
CAS  PubMed  Article  PubMed Central  Google Scholar 

42.
Lambert, G. I., Jennings, S., Kaiser, M. J., Hinz, H. & Hiddink, J. G. Quantification and prediction of the impact of fishing on epifaunal communities. Mar. Ecol. Prog. Ser. 430, 71–86 (2011).
ADS  Article  Google Scholar 

43.
Soldant, J., Mullier, T., Elliott, T. & Sheehan, E. V. Managing marine protected areas in Europe: Moving from ‘feature-based’ to ’whole-site; management of sites. In Marine Protected Areas: Science, Policy and Management Vol 828 (eds Humphreys, J. & Clark, R. W. E.) (Elsevier, New York, 2020).
Google Scholar 

44.
Staples, D. & Funge-Smith, S. Ecosystem Approach to Fisheries and Aquaculture: Implementing the FAO Code of Conduct for Responsible Fisheries (RAP Publication, Bangkok, 2009).
Google Scholar 

45.
Garcia, S. M., Rice, J. & Charles, A. Bridging fisheries management and biodiversity conservation norms: Potential and challenges ofbalancing harvest in ecosystem- based frameworks. Nature 6, 20 (2015).
Google Scholar 

46.
DEFRA. Marine Protected Areas Network Report 2012–2018. (2018).

47.
Burke, C. Ireland’s need for inshore local management. Fish. News 7, 20 (2015).
Google Scholar 

48.
Rees, S. E. et al. An evaluation Framework to Determine the Impact of the Lyme Bay Fisheries and Conservation Reserve and the Activities of the Lyme Bay Consultative Committee on Ecosystem Services and Human Wellbeing Final Report To the October 2016. (2016).

49.
Cork, M., McNulty, S. & Gaches, P. Site Selection Report for Inshore Marine SACs Project. Poole Bay to Lyme Bay. Report No. 9S0282/SSR/PooleLymeBay/01 (2008).

50.
Attrill, M. J. et al. Lyme Bay—A Case Study: Measuring Recovery of Benthic Species, Assessing Potential Spill-Over Effects and Socio-economic chaNges. (2012).

51.
Ross, R. South Devon Reef Video Baseline Surveys for the Prawle Point to Plymouth Sound & Eddystone cSAC and Surrounding Areas As commissioned by Natural England South Devon Reef Video Baseline Surveys for the Prawle Point to Plymouth Sound & Eddystone cSAC and Su. (2016)https://doi.org/10.13140/2.1.2313.1205.

52.
Vanstaen, K. & Eggleton, J. Mapping Annex 1 Reef Habitat Present in Specific areas Within the Lyme Bay and Torbay cSAC. (2011).

53.
Sheehan, E. V., Stevens, T. F., Attrill, M. J., Ropert-Coudert, Y. A quantitative, non-destructive methodology for habitat characterisation and benthic monitoring at offshore renewable energy developments. PLoS ONE 5(12), e14461 (2010).
ADS  CAS  PubMed  PubMed Central  Article  Google Scholar 

54.
Sheehan, E. V. et al. An experimental comparison of three towed underwater video systems using species metrics, benthic impact and performance. Methods Ecol. Evol. 7(7), 843–852 (2016).
Article  Google Scholar 

55.
Bicknell, A. W. J., Sheehan, E. V., Godley, B. J., Doherty, P. D. & Witt, M. J. Assessing the impact of introduced infrastructure at sea with cameras: A case study for spatial scale, time and statistical power. Mar. Environ. Res. 147, 126–137 (2019).
CAS  PubMed  Article  PubMed Central  Google Scholar 

56.
Priede, I. G., Bagley, P. M., Smith, A., Creasey, S. & Merrett, N. R. Scavenging deep demersal fishes of the porcupine seabight, North-East Atlantic: Observations by baited camera, trap and trawl. Nat. Hist. https://doi.org/10.1017/s0025315400047615 (1994).
Article  Google Scholar 

57.
Watson, D. L., Harvey, E. S., Anderson, M. J. & Kendrick, G. A. A comparison of temperate reef fish assemblages recorded by three underwater stereo-video techniques. Mar. Biol. 148, 415–425 (2005).
Article  Google Scholar 

58.
Cappo, M., Harvey, E. & Shortis, M. Counting and measuring fish with baited video techniques—an overview. Aust. Soc. Fish Biol. 1100, 1–9 (2006).
Google Scholar 

59.
Elliott, S. A. M., Turrell, W. R., Heath, M. R. & Bailey, D. M. Juvenile gadoid habitat and ontogenetic shift observations using stereo-video baited cameras. Mar. Ecol. Prog. Ser. 568, 123–135 (2017).
ADS  Article  Google Scholar 

60.
McLean, D. L., Harvey, E. S., Fairclough, D. V. & Newman, S. J. Large decline in the abundance of a targeted tropical lethrinid in areas open and closed to fishing. Mar. Ecol. Prog. Ser. 418, 189–199 (2010).
ADS  Article  Google Scholar 

61.
Harvey, E. S. et al. Comparison of the relative efficiencies of stereo-BRUVs and traps for sampling tropical continental shelf demersal fishes. Fish. Res. 125–126, 108–120 (2012).
Article  Google Scholar 

62.
Maragos, J. E. Marine and Coastal Biodiversity in the Tropical Island Pacific Region. (East-West Center, 1995).

63.
Clarke, K. R. & Warwick, R. M. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation (PRIMER-E Ltd., Plymouth Marine Laboratory, Plymouth, 2001).
Google Scholar 

64.
Taylor, P., Anderson, M. & Ter Braak, C. J. Stat. Comput. Permut. Tests Multi-Factor. Anal. Variance https://doi.org/10.1080/00949650215733 (2006).
Article  Google Scholar 

65.
Anderson, M. J. A new method for non-parametric multivariate analysis of variance MARTI. Austral Ecol. 26, 32–46 (2001).
Google Scholar  More

1.
Prokopy, R. J. & Roitberg, B. D. Joining and avoidance behavior in nonsocial insects. Annu. Rev. Entomol. 46, 631–665 (2001).
CAS  PubMed  Article  Google Scholar 
2.
Rudolf, V. H. & Rödel, M. O. Oviposition site selection in a complex and variable environment: the role of habitat quality and conspecific cues. Oecologia 142, 316–325 (2005).
ADS  PubMed  Article  PubMed Central  Google Scholar 

3.
Carrasco, D., Larsson, M. C. & Anderson, P. Insect host plant selection in complex environments. Curr. Opin. Insect Sci. 8, 1–7 (2015).
PubMed  Article  PubMed Central  Google Scholar 

4.
Dall, S. R., Giraldeau, L. A., Olsson, O., McNamara, J. M. & Stephens, D. W. Information and its use by animals in evolutionary ecology. Trends Ecol. Evol. 20, 187–193 (2005).
PubMed  Article  PubMed Central  Google Scholar 

5.
Kennedy, G. G. & Storer, N. P. Life systems of polyphagous arthropod pests in temporally unstable cropping systems. Annu. Rev. Entomol. 45, 467–493 (2000).
CAS  PubMed  Article  PubMed Central  Google Scholar 

6.
Prokopy, R. J. Marking pheromones. In: Capinera J. L. (eds) Encyclopedia of Entomology. Springer, Dordrecht. https://doi.org/10.1007/978-1-4020-6359-6_1730 (2008).

7.
Edmunds, A. J., Aluja, M., Diaz-Fleischer, F., Patrian, B. & Hagmann, L. Host marking pheromone (HMP) in the Mexican fruit fly Anastrepha ludens. CHIMIA Int. J. Chem. 64, 37–42 (2010).
CAS  Article  Google Scholar 

8.
Hauser, M. A historic account of the invasion of Drosophila suzukii (Matsumura) (Diptera: Drosophilidae) in the continental United States, with remarks on their identification. Pest Manag. Sci. 67, 1352–1357 (2011).
CAS  PubMed  Article  PubMed Central  Google Scholar 

9.
Calabria, G., Máca, J., Bächli, G., Serra, L. & Pascual, M. First records of the potential pest species Drosophila suzukii (Diptera: Drosophilidae) in Europe. J. Appl. Entomol. 136, 139–147 (2012).
Article  Google Scholar 

10.
Deprá, M., Poppe, J. L., Schmitz, H. J., De Toni, D. C. & Valente, V. L. The first records of the invasive pest Drosophila suzukii in the South American continent. J. Pest Sci. 87, 379–383 (2014).
Article  Google Scholar 

11.
Hassani, I. M. et al. First occurrence of the pest Drosophila suzukii (Diptera: Drosophilidae) in the Comoros Archipelago (Western Indian Ocean). Afr. Entomol. 28, 78–83 (2020).
Article  Google Scholar 

12.
Bellamy, D. E., Sisterson, M. S. & Walse, S. S. Quantifying host potentials: indexing postharvest fresh fruits for spotted wing drosophila, Drosophila suzukii. PLoS One, 8, e61227. https://doi.org/10.1371/journal.pone.0061227 (2013).

13.
Lee, J. C. et al. Infestation of wild and ornamental noncrop fruits by Drosophila suzukii (Diptera: Drosophilidae). Ann. Entomol. Soc. Am. 108, 117–129 (2015).
Article  Google Scholar 

14.
Kenis, M. et al. Non-crop plants used as hosts by Drosophila suzukii in Europe. J. Pest Sci. 89, 735–748 (2016).
Article  Google Scholar 

15.
Elsensohn, J. E. & Loeb, G. M. Non-crop host sampling yields insights into small-scale population dynamics of Drosophila suzukii (Matsumura). Insects 9, 5. https://doi.org/10.3390/insects9010005 (2018).
Article  PubMed  PubMed Central  Google Scholar 

16.
Mitsui, H., Takahashi, K. H. & Kimura, M. T. Spatial distributions and clutch sizes of Drosophila species ovipositing on cherry fruits of different stages. Pop. Ecol. 48, 233–237 (2006).
Article  Google Scholar 

17.
Atallah, J., Teixeira, L., Salazar, R., Zaragoza, G. & Kopp, A. The making of a pest: the evolution of a fruit-penetrating ovipositor in Drosophila suzukii and related species. Proc. R. Soc. B Biol. Sci. 281, 20132840. https://doi.org/10.1098/rspb.2013.2840 (2014).
Article  Google Scholar 

18.
Crava, C. M. et al. Structural and transcriptional evidence of mechanotransduction in the Drosophila suzukii ovipositor. J. Insect Physiol. 125, 104088 (2020).
CAS  PubMed  Article  Google Scholar 

19.
Burrack, H. J., Fernandez, G. E., Spivey, T. & Kraus, D. A. Variation in selection and utilization of host crops in the field and laboratory by Drosophila suzukii Matsumara (Diptera: Drosophilidae), an invasive frugivore. Pest Manag. Sci. 69, 1173–1180 (2013).
CAS  PubMed  Article  Google Scholar 

20.
Karageorgi, M. et al. Evolution of multiple sensory systems drives novel egg-laying behavior in the fruit pest Drosophila suzukii. Curr. Biol. 27, 847–853 (2017).
CAS  PubMed  PubMed Central  Article  Google Scholar 

21.
Silva-Soares, N. F., Nogueira-Alves, A., Beldade, P. & Mirth, C. K. Adaptation to new nutritional environments: larval performance, foraging decisions, and adult oviposition choices in Drosophila suzukii. BMC Ecol. 17, 21 (2017).
PubMed  PubMed Central  Article  CAS  Google Scholar 

22.
Olazcuaga, L. et al. Oviposition preference and larval performance of Drosophila suzukii (Diptera: Drosophilidae), spotted-wing Drosophila: Effects of fruit identity and composition. Environ. Entomol. 48, 867–881 (2019).
PubMed  Article  Google Scholar 

23.
Rendon, D. et al. Interactions among morphotype, nutrition, and temperature impact fitness of an invasive fly. Ecol. Evol. 9, 2615–2628 (2019).
PubMed  PubMed Central  Article  Google Scholar 

24.
Scheidler, N. H., Liu, C., Hamby, K. A., Zalom, F. G. & Syed, Z. Volatile codes: correlation of olfactory signals and reception in Drosophila-yeast chemical communication. Sci. Rep. 5, 14059. https://doi.org/10.1038/srep14059 (2015).
ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

25.
Keesey, I. W. et al. Adult frass provides a pheromone signature for Drosophila feeding and aggregation. J. Chem. Ecol. 42, 739–747 (2016).
CAS  PubMed  PubMed Central  Article  Google Scholar 

26.
Lasa, R., Navarro-de-la-Fuente, L., Gschaedler-Mathis, A. C., Kirchmayr, M. R. & Williams, T. Yeast species, strains, and growth media mediate attraction of Drosophila suzukii (Diptera: Drosophilidae). Insects 10, 228. https://doi.org/10.3390/insects10080228 (2019).
Article  PubMed Central  PubMed  Google Scholar 

27.
Bellutti, N. et al. Dietary yeast affects preference and performance in Drosophila suzukii. J. Pest Sci. 91, 651–660 (2018).
Article  Google Scholar 

28.
Wallingford, A. K., Hesler, S. P., Cha, D. H. & Loeb, G. M. Behavioral response of spotted-wing drosophila, Drosophila suzukii Matsumura, to aversive odors and a potential oviposition deterrent in the field. Pest Manag. Sci. 72, 701–706 (2016).
CAS  PubMed  Article  Google Scholar 

29.
Tait, G. et al. Reproductive site selection: evidence of an oviposition cue in a highly adaptive dipteran, Drosophila suzukii (Diptera: Drosophilidae). Environ. Entomol. 49, 355–363 (2020).
CAS  PubMed  Article  Google Scholar 

30.
Lin, C. C., Prokop-Prigge, K. A., Preti, G. & Potter, C. J. Food odors trigger Drosophila males to deposit a pheromone that guides aggregation and female oviposition decisions. Elife 4, e08688. https://doi.org/10.7554/eLife.08688 (2015).
Article  PubMed  PubMed Central  Google Scholar 

31.
Duménil, C. et al. Pheromonal cues deposited by mated females convey social information about egg-laying sites in Drosophila melanogaster. J. Chem. Ecol. 42, 259–269 (2016).
PubMed  PubMed Central  Article  CAS  Google Scholar 

32.
Barker, J. S. F. & Podger, R. N. Interspecific competition between Drosophila melanogaster and Drosophila simulans: effects of larval density on viability, developmental period and adult body weight. Ecol. 51, 170–189 (1970).
Article  Google Scholar 

33.
Rohlfs, M., Obmann, B. & Petersen, R. Competition with filamentous fungi and its implication for a gregarious lifestyle in insects living on ephemeral resources. Ecol. Entomol. 30, 556–563 (2005).
Article  Google Scholar 

34.
Durisko, Z., Anderson, B. & Dukas, R. Adult fruit fly attraction to larvae biases experience and mediates social learning. J. Exp. Biol. 217, 1193–1197 (2014).
PubMed  Article  Google Scholar 

35.
Prokopy, R. J. & Duan, J. J. Socially facilitated egglaying behavior in Mediterranean fruit flies. Behav. Ecol. Sociobiol. 42, 117–122 (1998).
Article  Google Scholar 

36.
Elsensohn, J. Factors affecting oviposition behavior. In Drosophila suzukii (North Carolina State University, Raleigh, 2020).

37.
Hardin, J. A., Kraus, D. A. & Burrack, H. J. Diet quality mitigates intraspecific larval competition in Drosophila suzukii. Entomol. Exp. Appl. 156, 59–65 (2015).
CAS  Article  Google Scholar 

38.
Averill, A. L. & Prokopy, R. J. Intraspecific competition in the tephritid fruit fly Rhagoletis pomonella. Ecol. 68, 878–886 (1987).
Article  Google Scholar 

39.
Arredondo, J. & Diaz-Fleischer, F. Oviposition deterrents for the Mediterranean fruit fly, Ceratitis capitata (Diptera: Tephritidae) from fly faeces extracts. Bull. Entomol. Res. 96, 35–42 (2006).
CAS  PubMed  Article  Google Scholar 

40.
Nufio, C. R. & Papaj, D. R. Host marking behavior in phytophagous insects and parasitoids. Entomol. Exp. Appl. 99, 273–293 (2001).
Article  Google Scholar 

41.
Papaj, D. R. Use and avoidance of occupied hosts as a dynamic process in tephritid flies. In Insect-Plant Interactions (ed. Bernays, E.A.) 25–46 (CRC Press, Boca Raton, 2017).

42.
Bernays, E.A. & Chapman, R.F. Behavior: the process of host-plant selection. In Host-Plant Selection by Phytophagous Insects Vol. 2. (eds. Bernays, E.A. & Chapman, R.F) 95–165 (Springer Science & Business Media, Berlin, 2007).

43.
Schoonhoven, L. M. Host-marking pheromones in Lepidoptera, with special reference to two Pieris spp. J. Chem. Ecol. 16, 3043–3052 (1990).
CAS  PubMed  Article  PubMed Central  Google Scholar 

44.
Dancau, T., Stemberger, T. L., Clarke, P. & Gillespie, D. R. Can competition be superior to parasitism for biological control? The case of spotted wing Drosophila (Drosophila suzukii), Drosophila melanogaster and Pachycrepoideus vindemmiae. Biocontrol Sci. Tech. 27, 3–16 (2017).
Article  Google Scholar 

45.
Rosenheim, J. A. The relative contributions of time and eggs to the cost of reproduction. Evol. 53, 376–385 (1999).
Article  Google Scholar 

46.
Jiménez-Padilla, Y., Ferguson, L. V. & Sinclair, B. J. Comparing apples and oranges (and blueberries and grapes): fruit type affects development and cold susceptibility of immature Drosophila suzukii (Diptera: Drosophilidae). Can. Entomol. 152, 532–545 (2020).
Article  Google Scholar 

47.
Papaj, D. R. & Messing, R. H. Functional shifts in the use of parasitized hosts by a tephritid fly: the role of host quality. Behav. Ecol. 7, 235–242 (1996).
Article  Google Scholar 

48.
Ingleby, F. C. Insect cuticular hydrocarbons as dynamic traits in sexual communication. Insects 6, 732–742. https://doi.org/10.3390/insects6030732 (2015).
Article  PubMed  PubMed Central  Google Scholar 

49.
Snellings, Y. et al. The role of cuticular hydrocarbons in mate recognition in Drosophila suzukii. Sci. Rep. 8, 1–11 (2018).
CAS  Article  Google Scholar 

50.
Li, G. & Ishikawa, Y. Oviposition deterrents in larval frass of four Ostrinia species fed on an artificial diet. J. Chem Ecol. 30, 1445–1456 (2004).
CAS  PubMed  Article  Google Scholar 

51.
Wada-Katsumata, A. et al. Gut bacteria mediate aggregation in the German cockroach. Proc. Nat. Acad. Sci. 112, 15678–15683 (2015).
ADS  CAS  PubMed  Article  Google Scholar 

52.
Mansourian, S. et al. Fecal-derived phenol induces egg-laying aversion in Drosophila. Curr. Biol. 26, 2762–2769 (2016).
CAS  PubMed  Article  Google Scholar 

53.
Bueno, E. et al. Response of wild spotted Wing Drosophila (Drosophila suzukii) to microbial volatiles. J. Chem. Ecol. 39, 1–11 (2019).
Google Scholar 

54.
Behar, A., Jurkevitch, E. & Yuval, B. Bringing back the fruit into fruit fly–bacteria interactions. Mol. Ecol. 17, 1375–1386 (2008).
CAS  PubMed  Article  Google Scholar 

55.
Yun, J. H. et al. Insect gut bacterial diversity determined by environmental habitat, diet, developmental stage, and phylogeny of host. Appl. Environ. Microbiol. 80, 5254–5264 (2014).
PubMed  PubMed Central  Article  CAS  Google Scholar 

56.
Bing, X., Gerlach, J., Loeb, G. & Buchon, N. Nutrient-dependent impact of microbes on Drosophila suzukii development. MBio 9, e02199-e2217. https://doi.org/10.1128/mBio.02199-17 (2018).
CAS  Article  PubMed  PubMed Central  Google Scholar 

57.
Martinez-Sañudo, I. et al. Metagenomic analysis reveals changes of the Drosophila suzukii microbiota in the newly colonized regions. Insect Sci. 25, 833–846 (2018).
PubMed  Article  Google Scholar 

58.
Silva, M. A., Bezerra-Silva, G. C. D. & Mastrangelo, T. The host marking pheromone application on the management of fruit flies—a review. Braz. Arch. Biol. Tech. 55, 835–842 (2012).
CAS  Article  Google Scholar 

59.
Hamby, K. A. & Becher, P. G. Current knowledge of interactions between Drosophila suzukii and microbes, and their potential utility for pest management. J. Pest Sci. 89, 621–630 (2016).
Article  Google Scholar 

60.
Alkema, J. T., Dicke, M. & Wertheim, B. Context-dependence and the development of push-pull approaches for integrated management of Drosophila suzukii. Insects 10, 454. https://doi.org/10.3390/insects10120454 (2019).
Article  PubMed  PubMed Central  Google Scholar 

61.
Revadi, S. et al. Sexual behavior of Drosophila suzukii. Insects 6(183), 196. https://doi.org/10.3390/insects6010183 (2015).
Article  Google Scholar 

62.
Elya, C. et al. Robust manipulation of the behavior of Drosophila melanogaster by a fungal pathogen in the laboratory. Elife 7, e34414. https://doi.org/10.7554/eLife.34414 (2018).
Article  PubMed  PubMed Central  Google Scholar 

63.
Emiljanowicz, L. M., Ryan, G. D., Langille, A. & Newman, J. Development, reproductive output and population growth of the fruit fly pest Drosophila suzukii (Diptera: Drosophilidae) on artificial diet. J Econ. Entomol. 107, 1392–1398 (2014).
PubMed  Article  Google Scholar 

64.
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ (2019). More