Ecology

Environmental isotopic baseline of the Yingpan cemeteryMultiple body tissues of Yingpan Man, including his serial dentine, bone collagen, hair keratin and muscle, were isotopically analyzed. The δ13C measurements show a mix of C3 and C4 dietary inputs with values that range from − 17.3‰ to − 13.7‰, while the δ15N results are remarkably elevated and range from 14.9‰ to 19.9‰. According to the isotopic literature, plant δ15N values above 20‰ are relatively rare in terrestrial environments22,26,27,28. However, the δ15N values of the plants, animals and Yingpan Man are remarkably elevated (up to 23.3‰) compared to isotopic results from other parts of China and Europe (Fig. 3)12. This suggests that the entire ecosystem of this portion of Xinjiang is 15N-enriched. In addition, seeds of Nitraria pamirica having elevated δ15N results (up to 27.9‰) were found in the western region of the Pamir Plateau near the Afghanistan border23. Grains of wheat and barley from Shichengzi (ca. 202 BC–220 AD; ~ 400 km from Yingpan) yielded δ15N values that range from 14.6‰ to 19.8‰ (mean ± SD = 17.2 ± 1.5‰; n = 10) and 13.4‰ to 19.8‰ (mean ± SD = 17.3 ± 1.9‰; n = 12), respectively29. This evidence suggests that the entire Tarim Basin and greater Xinjiang has some of the most elevated terrestrial δ15N values in Eurasia. This is also supported by past isotopic work in China, which demonstrated a correlation between human δ15N values and annual mean precipitation, with individuals from Xinjiang having the highest δ15N values of all regions studied12. The cause of this 15N-enrichment must be at least partially environmental based on past isotopic studies27,30. The extreme aridity of the Yingpan cemetery site, which is located in the Taklamakan Desert, the driest region of China and characterized by little rainfall of 0 to 100 mm/yr and high evapotranspiration rates  > 2500 mm/year (Fig. 1)31, results in a large plant 15N-enrichment by intensive evaporation of 15N-depleted ammonia (NH4+) from the soil. These elevated δ15N results are then translated up the food chain to the domestic animals and humans. This is compelling evidence that Yingpan man was born and raised in a 15N-enriched environment that was extremely arid, and that he consumed wheat/barley, broomcorn millet and grape that were grown locally.Breastfeeding, weaning and childhood dietary patterns of Yingpan ManIsotopic analysis of dentine serial sections in human teeth permit an investigation of individual dietary patterns over the time period when the teeth were developing24. For Yingpan Man, his first molar (M1), represents the period of his life from birth to approximately 10 years old32. The first five 1 mm dentine sections from the M1 crown, corresponding to the first ~ 2.2 years of his life, show a steady decrease in δ15N of 2.5‰ (Fig. 4). This is evidence that Yingpan Man was fully weaned off breastmilk at or before the age of ~ 2.2 years old (Fig. 4)33,34, and is similar to findings of the Zhou Dynasty (1122–771 BC) sites of Boyangcheng35, Xiyasi and Changxinyuan36.Archaeological evidence for infant feeding practices is rarely preserved in Xinjiang. However, two extraordinary feeding vessels (one made of goat breast skin, the other made from an ox horn) were previously discovered with the burial of a 10-month old infant mummy in the southern Tarim Basin4,7. This infant dates to ~ 1000 BC and was found at the Zhagunluke site which is ~ 460 km from the Yingpan cemetery (Fig. 1). In addition, historical documents also provide supporting evidence that children were fed with animal milk in ancient Xinjiang. For example, the Kharosthī scripts (written in a version of the north Indian Prakrit language) are a collection of contracts, letters and other documents (e.g. wood tablets, leather, silk, paper, etc.) that detail life, events, trading, taxes and agricultural practices during the third to fifth century AD of the Shanshan Kingdom (~ 250 km from Yingpan, see Fig. 6)37. It is recorded in these documents that a “milk fee” of cattle or camel was paid to the adoptee when children were adopted in ancient Shanshan, and this exchange was not only protected by the law but also needed to be ensured by witnesses37,38. Thus, there may have been a common tradition of feeding infants with milk from domestic animals (e.g. goat, cattle, camel, etc.) during the weaning process in ancient Xinjiang.Figure 6Yingpan Man’s grave goods mirroring the cultural contact across the Eurasian continent (~ 300 AD). (a) Decorative patterns of vase and flowers on the sideboards of Yingpan Man’s coffin. (b) Decorative patterns of nude putties, goats, bulls and pomegranate trees on Yingpan Man’s woolen caftan. (c) Yingpan Man’s white hemp mask with golden diadem (possibly related to the death masks from Tashtyk culture). (d) Woolen trousers with decorative patterns of double quatrefoil florals surrounded by lozenges made up of solid circles and flowers. (e) Miniature funerary summer coat. (f) Miniature funerary winter robe. (g) Embroidered armband. (h) Yellow painted wood coffin with decorative patterns. (i) A broken brocade that was decorated with Chinese characters of “Shou” and “You”. (j) Silk fish. (k) Crowing cockerel pillow. (l) Perfume sachet. (m) A tufted carpet with decorative patterns of a male lion. (Map was generated using GMT 5.2.1. The original pictures of artifacts were previously published14 and provided by Wenying Li. Pictures were modified using Adobe Photoshop CC 2015 V.1.2. The final layout was created in Adobe Illustrator CC 2019 V.23.1.1).Full size imageIn contrast, the δ13C results of the first four dentine sections show little variability, which indicates that Yingpan Man consumed other foods/liquids from an early age and was not exclusively breastfed for a significant amount of time after birth (Fig. 4; Fig. S4)34. Between 4 and 9 mm, there is an increase in the δ13C values of 1.7‰, which corresponds to a late weaning and childhood diet from approximately 1.8 to 4.2 years old, where more millet was consumed (Fig. 4; Fig. S4). This millet possibly took the form of a gruel as there is little evidence that Yingpan Man consumed increased amounts of animal protein during this period. Similar isotopic patterns in dentine representing individuals that consumed increasing amounts of millet during weaning and early childhood were found at the Late Neolithic (4500 BP) Gaoshan site in Sichuan Province39. This suggests that millet may have had a long history of being used as a weaning or childhood food in China. From 9 to 14 mm, the dentine serial sections show a decline in δ13C values of 1.4‰ which indicates a period of increasing C3 foods in the diet from approximately 4.2 to 6.6 years old (Fig. 4; Fig. S4). Then from 14 to 21 mm the δ13C values increase again by 1.3‰, evidence of a change back to millet consumption from approximately 6.6 to 10 years old. Thus, over the first 10 years of life, Yingpan Man had frequent dietary shifts between C3-based and C4-based foods (Fig. 4; Fig. S4).Seasonal diet of Yingpan ManIn agreement with his dentine serial sections, the δ13C results of Yingpan Man’s hair display fluctuations between C3-based (e.g. wheat, barley) and C4-based (millet) foods (Fig. 5a). These hair δ13C results appear to follow a periodic trend, and suggest that his diet changed during different seasons of the year. In contrast, apart from the last 6–7 months of life, Yingpan Man’s hair δ15N values show little variation (Fig. 5b). However, a strong correlation was found between the variability of both the hair δ13C and δ15N values (Sperman’s r = − 0.534; sig = 0; n = 46; Fig. S5a), but not between the δ13C and δ34S values (Sperman’s r = 0.100; sig = 0.539; n = 40; Fig. 5; Fig. S5b). The lack of significant changes in the δ15N values is evidence that Yingpan Man’s protein consumption was relatively constant for approximately the last 4 years before death. This constant protein consumption, but variable intake of C3 and C4 plants, could suggest that in addition to directly consuming different amounts of wheat/barley and millet, Yingpan Man may have actually consumed domestic animals (goats, sheep, cattle) that were foddered on these crops at different times of the year. This possibility could explain the periodic variability in the δ13C values as well as the lack of variability in the δ15N values. In addition, the δ34S hair results also show little variation in the serial sections (~ 1.5‰). Yingpan Man’s δ34S results have a terrestrial range between 10.4‰ and 11.9‰40,41,42 and are similar to past δ34S values from the Proto-Shang site of Nancheng in Hebei Province43. While few δ34S studies have been reported for China for comparison, this lack of isotopic variability in sulfur suggests that Yingpan Man was likely not a Silk Road traveler, but stayed close to the local area during the last years of his life42. Future research involving strontium analysis on Yingpan Man’s hair serial sections and teeth will hopefully support or refute these findings44.Regional evidence in support of seasonal dietary changes comes from apatite δ13C and δ18O results of serial sections of tooth enamel of domesticates from the pastoral sites of Dali, Begash and Tasbas in Kazakhstan (~ 750 km from Yingpan)45. Specifically, data from early and middle Bronze Age sheep, goats and cattle displayed periodic patterns in their δ13Capa and δ18Oapa results that reflect the consumption of more millet during winter and more C3 plants during summer months, based on environmental inputs of body water45. A combination of radiocarbon dates and the application of the dietary mixing model (MixSIAR) identified early sheep and goats from this region (e.g. Dali, 2705–2545 cal BC) to be winter foddered with up to 44–50% of millet intake45. During later periods, this reliance on millet fodder during the winter months increased with some goats from Begash having 50–60% millet in their diets. Further, the relative contribution of millet to the diet of sheep and goats from the early phase (2345–2080 cal BC) of Begash reached up to 68–74% of the whole diet during winter months, especially from November to December45. This key study indicates that there is a long history and precedent for seasonal C3 and C4 feeding of domestic animals in Central Asia, and supports the isotopic findings of a seasonal diet in the hair of Yingpan Man.In addition, isotopic analysis of sequential hair samples from mummies recovered from the Oglakhty cemetery in the Minusinsk Basin of southern Siberia, Russia (~ 900 km north of Yingpan) also show seasonal dietary variation with millet and fish consumed during the late summer and autumn46. These mummies of the Tashtyk culture date to the same period as Yingpan Man (third to fourth centuries AD), and interestingly were also buried with white painted gypsum funerary masks that are similar to the one that Yingpan Man was wearing46,47. This unique burial tradition could suggest links such as trade or an association between Yingpan Man and the Tashtyk culture and additional research is necessary to explore this possibility in more detail (Fig. 6).Historical documents such as the Kharosthī scripts provide additional valuable information about seasonal diets and the foods consumed by the inhabitants of ancient Xinjiang37. These texts describe how the people of the Tarim Basin cultivated mainly wheat, millet and barley as their main cereal crops and that grapes were carefully managed for the production of wine37. Autumn (around the 10th month of the year), was mentioned as the time for harvesting crops and trading crops as well as wine and animals, and paying debts37. Thus, autumn would have been the time of the year with the most abundant amount of food resources, especially C3 foods like fruit, vegetables and wine. This is still true today, as the harvesting of agricultural products in the Tarim Basin mainly occurs in the month of October48. In contrast, during the winter months food resources would have been scarce with only non-perishable crops like millets and wheat available to guarantee the food supply “in the harsh winters of Inner Asia”45. Thus, more C4 foods were likely consumed during the winter and spring months while more C3 foods were consumed during the summer and autumn months.If this information is applied to Yingpan Man’s hair δ13C results, it would suggest that the δ13C values decreased during the summer and autumn months (June to October) and likely reach a nadir during the middle of autumn or October. Therefore, Yingpan Man’s hair sections which are 8 to 10 cm, 18 to 25 cm and 36 to 43 cm from his scalp reflect the period of middle autumn (between August to October) while the hair sections which are 4 to 6 cm, 13 to 15 cm and 29 to 33 cm from his scalp represent late winter (December to February). This would suggest that Yingpan Man died ~ 4 months after the last drop in his δ13C hair values or in spring, possibly March or April (Fig. 5). In addition, the clothes in which Yingpan Man was buried, as well as his associated wardrobe, also provide information about the timing of his death14,15. The miniature robe placed on his chest was designed for winter, as it was long, double layered and the interior was lined with sheep’s wool (Fig. 2f). Whereas the miniature coat placed near his wrist was designed for summer as it was shorter and made only of a single layer of silk (Fig. 2p). Yingpan Man’s caftan, in which he was buried, was double layered with the outer layer made of wool and the inner lining made of silk, and this was likely designed either for the spring or autumn (Fig. 2k)14. Thus, Yingpan Man’s burial clothes combined with the isotopic and historical evidence indicate that Yingpan Man died sometime during the spring months14,15.To better illustrate Yingpan Man’s seasonal dietary variation, two hair sections respectively representing the highest (Sample A: − 14.0 ± 0.1‰, 15.2 ± 0.1‰) and lowest δ13C values (Sample B: − 17.3 ± 0.2‰, 15.8 ± 0.1‰) were analyzed by a Bayesian mixing model with the application of FRUITS (Food Reconstruction Using Isotopic Transferred signals)49,50. As displayed in Fig. S6, isotopic data for millet, wheat/barley, grape and sheep/goat from the Yingpan cemetery were incorporated into the mixing model as feasible dietary sources for Yingpan Man in both scenarios (Sample A and Sample B). The isotopic fractionation between human hair and diet is corrected with an offset of 4.0 ± 0.5‰ for δ13C values51,52,53 and 4.5 ± 0.5‰ for δ15N values51,54. The relative contribution of different macronutrients is defined in the mixing model according to published records with 74 ± 4% of the carbon originating from protein and 26 ± 4% originating from carbohydrates and lipids and all of the nitrogen originating from protein55,56. The high hair δ13C value of − 14.0‰ (Sample A) represents a heavy reliance on millet consumption (39–67%; median = 53%) and a low amount of dietary input from wheat/barley (0–51%; median = 12%), grape (1–46%; median = 25%) and sheep/goat (0–30%; median = 4%), which likely reflects Yingpan Man’s diet during the winter and spring months (Fig. S6 and Table S13). In contrast, the low hair δ13C value of − 17.3‰ (Sample B) likely represents a summer/autumn diet with a decline in C4 foods, like millet (8–44%; median = 28%), as well as increased reliance on C3 foods like wheat/barley (1–71%; median = 21%) and sheep/goat (0–81%; median = 9%) (Fig. S6 and Table S13). However, the importance of sheep/goat and wheat/barley is likely under-estimated here given their small sample size and highly elevated δ15N values. Nonetheless, the FRUITS mixing model indicates that the varying consumption of plant foods, especially millet, is clearly responsible for the δ13C shifts of Yingpan Man’s diet.The last months of Yingpan ManThe hair δ13C results provide an estimate for the time of year Yingpan Man likely died. In addition, the hair δ15N and δ34S values can provide evidence for how Yingpan Man may have died. The last ~ 6 cm (closest to the scalp) of Yingpan Man’s hair show a general rise in δ15N by ~ 1‰ (Fig. 5b). This pattern is uncharacteristic compared to the other hair δ15N results that display little variation. This unique 15N-enrichment could represent a period of catabolic wasting due to the recycling of tissue proteins as a result of a prolonged illness25,57,58. Additional support for some sort of disease or period of illness comes from the fact that there is little change in the δ13C values but a slight decrease in the δ34S values during the last ~ 6 cm of Yingpan Man’s hair. Tissue catabolism is known to cause an increase in δ15N but little change in δ13C values in human hair25. Further, δ34S results are known to decrease in the red blood cells and serum (by ~ 1.5‰) of patients suffering from liver cancer59. As the last ~ 6 cm of hair displayed a slight decrease in δ34S by ~ 1‰, with the largest decline during the last month of life, this evidence in conjunction with the δ13C and δ15N values suggests that Yingpan Man did not die suddenly but likely suffered some type of debilitating sickness over the last months of his life before he succumbed. However, as most wasting diseases or illnesses leave no traces on human skeletons, it is difficult to define the specific disease that caused Yingpan Man’s death, and a detailed paleo-pathological study of Yingpan Man is needed in the future.The grave goods buried with Yingpan Man also suggest he may have suffered a compromised health status before death. In particular, a piece of tattered yellow brocade decorated with brown and blue images of vines, animals, birds, flowers, as well as the Chinese characters of “Shou” and “You” was found placed at a prominent position on the right side of Yingpan Man’s head (Fig. 7)14. According to Chinese historical literature sources, “You” means “blessing” or “blessed”. While, “Shou” means “long live” or “healthy” (Fig. 7)60. Similar grave goods with Chinese characters e.g. “Yan Nian Yi Shou Da Yi Zi Sun” (meaning “live longer and benefit the descendants”), “Chang Le Ming Guang Cheng Fu Shou You” (meaning “always be happy, be bright, be lucky and be blessed”), “Yong Chang” (meaning “always be prosperous”), were also frequently unearthed from contemporary and later cemeteries of the Tarim Basin, especially the nearby sites of the Loulan culture (third century BC to 448 AD, renamed as Shanshan in 77 BC, ~ 250 km from Yingpan), as grave goods that carry good wishes for the dead15. However, Yingpan Man’s brocade shows significant traces of wear or “rubbing” (Fig. 7). This is interesting as this brocade was not complete or new but well-worn to the point of being tattered and frayed, yet it was still placed at a very important position in Yingpan Man burial—just beside his head14. This suggests that this brocade may have been an important “lucky” health charm for Yingpan Man that was frequently used either by himself or by those who cared for him before burial. Thus, both isotopic and archaeological evidence suggest Yingpan Man suffered some type of illness during the last ~ 6 months of his life, likely in winter, and that he succumbed to this illness in the following spring.Figure 7Photo and sketch showing details of the embroidered brocade that was found with Yingpan Man burial. (The original picture and sketch were previously published14 and provided by Wenying Li. Pictures were modified using Adobe Photoshop CC 2015 V.1.2. The final layout was created in Adobe Illustrator CC 2019 V.23.1.1).Full size imageWho was Yingpan Man?The social identity and status of Yingpan Man is enigmatic given the various cultural components of his grave goods and since the collection of his funerary objects are unique compared to all of the other burials found in the same cemetery14. This has created much controversy and debate about the social identity of Yingpan Man7,14,15,16. Thus, “Who was Yingpan Man?” and “Why was he buried here, in a normal unmarked grave with such lavish and exotic grave goods?” is an active topic of debate.The physical anthropology of Yingpan Man was investigated but not formally published. According to which, the metric and nonmetric index of traits of his skeletal remains suggest a mix of both European and Mongolian features (Dong Wei, personal communication). However, facial reconstruction conducted by Dong Wei (personal communication) indicates that Yingpan Man’s facial structure is more characteristic of the features from Western Eurasia. This is consistent with the image of the face painted on his death mask and the fact that he wears a golden diadem across his forehead, which is more traditionally associated with Greece4,7. However, Yingpan Man’s white hemp mask is similar in style to the white painted gypsum masks of the Tashtyk culture from the Minusinsk Basin of Russia46. In addition, Yingpan Man and the mummies of the Oglakhty cemetery of Tashtyk date to nearly the same period (third to fourth centuries AD), and the polychrome silk cloth from the Tarim Basin has also been discovered in Oglakhty46. This evidence, while circumstantial, could suggest some form of association or that links with the Tashtyk may have taken place during the lifetime of Yingpan Man, possibly through trade or familial relationships (Fig. 6).Other components of Yingpan Man’s burial provide important clues about his social identity and cultural affiliations in life61. The styles and types of grave goods of Yingpan Man display an unique mix of both Eastern and Western cultures and traditions that were likely common to inhabitants of Silk Road trading towns in Xinjiang during the third to fourth centuries AD (Fig. 6)14,15,16. However, there appears to be a strong eastern component in some of his funerary arrangements. For example, the burial practices associated with Yingpan Man: covering his face, filling his nose, burying his body fully clothed, covering it with a silken burial shroud, as well as using miniature funeral objects as grave goods are in accordance with the suggested burial rites of “Yan” (meaning “covering”), “Zhen” (meaning “filling”), “She Min Mu” (meaning “covering the eyes”), “Qin” (meaning “quilt” or “burial shroud”) mentioned in the Confucian literature of Yili (meaning “Rites”, formed during the Zhou Dynasties (1046 BC to 256 BC))15,62. In addition, the styles and designs of some of his grave goods are indicative of Chinese spiritual beliefs. For instance, the diamond and circle-shaped designs of decorative patterns on the cover and sideboards of Yingpan Man’s coffin are argued by some scholars to be the traditional “Lianbi” pattern (meaning “linked jades”) which symbolizes the jade burial suits (“Jinlü Yuyi”) that were popular in Han Dynasty burials of high status nobles in central and southern China (e.g. Nanyuewang tomb, Mancheng tomb)15. In particular, as jades are believed to be the ideal material for embalming in ancient China63, the “Lianbi” pattern on Yingpan Man’s coffin carries the symbolic meaning of preserving his body forever so that his soul and spirt will reach heaven (Fig. 6; Fig. S1)15.The brocade “health charm” found to the right side of Yingpan Man also shows a clear affiliation with China, as it was decorated with the Chinese characters of “Shou” and “You”. This is important as some historical documents suggest that Kharosthī was the common language in the Tarim Basin at this time37,38, but Yingpan Man was likely more accustomed to Chinese spiritual beliefs according to this brocade (Figs. 6, 7). Moreover, the crowing cockerel pillow is also interesting as it was recorded in the Han Dynasty literature of Lunheng (meaning “On Balance”, compiled by Chong Wang during the Eastern Han Dynasty in 88 AD), that the dead “turns into ghosts” and “hurt the living ones”64, and that a crowing cockerel was believed to be able to expel evil spirts and ghosts65. Thus, a crowing cockerel pillow was frequently used in funerary practices in China since the Han Dynasty66, and it is still a common grave good today in some areas of modern China (e.g. Shandong and Guizhou Provinces)65.The eight decorative pearls attached to the crowing cockerel pillow are also important status markers regarding the identity of Yingpan Man (Fig. 8). In Xinjiang, pearls would have been long distance imported products from the coastal regions which are  > 3000 km from the Yingpan cemetery. Ancient China, Egypt, Persia, Greece and India were known to have produced and prized pearls67. However, according to Chinese historic literature sources: Shangshu (meaning “the Book of Documents”; written by pre-Qin philosophers during the Zhou Dynasty ~ 1000 BC)68, Hanshu (meaning “the Book of Han”; written by Gu Ban during the Eastern Han Dynasty in 105 AD)69 and Hou Hanshu (meaning “the Book of Later Han”; written by Ye Fan during the Southern Dynasties from 432 to 445 AD)70, pearls were known as royal tributes in China since the Xia Dynasty (~ 2000 BC), and were mainly produced in the coastal cities of southern China, such as Panyu (modern Guangzhou City in Guangdong Province), Hepu (modern Hainan Province), Zhuya (modern Guangxi and Guangdong Province), etc.67 These coastal regions of China were connected to the Tarim Basin area via the Silk Road trading routes and were possibly the source for the pearls found on Yingpan Man’s pillow (Fig. 6). In particular, a silk pillow fully covered with unpolished natural pearls (placed beneath the tomb owner’s head, weighing ~ 470 g) and a lacquer box of pearls (weighing 4,117 g) were unearthed from the tomb of the King of the Nanyue Kingdom (a vassal state of Han) in modern Guangzhou71. This demonstrates a similar preference for pearls as decorations on pillows or as grave goods for high status individuals.Figure 8Crowing cockerel pillow. (a–d) Photo and sketch of the crowing cockerel pillow from Yingpan Man burial. (e–h) Decorative patterns of the four mythical beasts on Han Dynasty eaves tiles. (The original pictures and sketches were previously published14 and provided by Wenying Li. Pictures were modified using Adobe Photoshop CC 2015 V.1.2. The final layout was created in Adobe Illustrator CC 2019 V.23.1.1).Full size imageMoreover, the decorative patterns found on Yingpan Man’s crowing cockerel pillow are remarkable. These include the images of: a monkey-shaped face, a griffin-shaped beast and a net-shaped pattern as well as the images of the mythical beasts of the “blue dragon”, “red sparrow” and “white tiger” (Fig. 8d)17,47,61,72. Together with another mythical beast known as the “black turtle-snake”, these four mythical beasts were believed to be the guardians of the four cardinal directions and also represents four different colors and elements according to traditional Chinese cosmology (Fig. 8e–h). The dragon guarding the East and representing the color blue and the element of wood, the sparrow guarding the South and representing the color red as well as the element of fire, the tiger guarding the West and representing the color white and the element of metal, and the turtle-snake guarding the North and representing the color black and the element of water62,73. Thus, the images of these four mythical beasts were frequently integrated into the designs of cities, buildings, cemeteries and objects of ancient China for their symbolic function of protection63. In this context, the net-shaped pattern on Yingpan Man’s crowing cockerel pillow is likely a symbol of the four cardinal directions, while the monkey shaped face possibly symbolizes a human, that is to be protected. Notably, here, in the design of Yingpan Man’s crowing cockerel pillow, the image of the northern guardian of the turtle-snake is replaced by the Greek mythical beast known as Griffin (Fig. 8d)16,47,74. This change in imagery and symbolism could suggest that contact between the East and the West was via a northern route and that this was associated with some type of Greek influence or the Griffin. According to historical literature sources, a large number of Greeks migrated into Central Asia with Alexander the Great’s eastward expedition3,5,6. Hellenistic kingdoms such as Bactria (a.k.a. Daxia in the Chinese literature, located in the area of Pamir Plateau in modern Afghanistan, ~ 1900 km from Yingpan) subsequently became centers of Greek influence in this area and exported elements of Hellenistic culture to surrounding kingdoms that were located along the main routes of the Silk Road, such as Dayuan (located in the Fergana Valley, ~ 1300 km from Yingpan)2,6,75. Some ancient cities in the Tarim Basin area, such as Yingpan, Maideke and Yuansha, are also argued to have been influenced by Hellenistic elements as their city walls were circular in shape and this was clearly a unique architecture compared to the square-shaped traditional Chinese city walls63,76. It is highly probably that these Hellenistic elements were transported to the ancient city of Yingpan by the Silk Road trading routes which followed a north–south direction in this region, and therefore would support that Greek elements would be associated with the northern direction in this region of Xinjiang. In conclusion, Yingpan Man’s crowing cockerel pillow displays an unprecedented combination of Chinese spiritual beliefs incorporated with western motifs, and symbolizes the intertwining of Eastern and Western cultures in this part of Central Asia (Figs. 6, 8).Even more significant uses of western components in Yingpan Man’s burial is also visualized in other grave goods, e.g. his: carpet, coffin, caftan. The decorative pattern of the male lion on Yingpan Man’s tufted woolen carpet is clearly an imported element as the lion was originally from Africa, Southern Europe, West Asia and India. Lions were not introduced into China until 87 AD, being sent to Emperor Zhang of the Eastern Han Dynasty as a gift from Pacorus II, the King of the Anxi Empire (a.k.a. Parthian Empire (247 BC to 224 AD, replaced by the Persian in 226 AD, located in modern Iran))70. In particular, this exquisite lion-decorated carpet suggests a very high social status for Yingpan Man as lions were deified in ancient China since the Han Dynasty and were used as symbols of power, authority and royalty as it was in other areas of the Eurasia (Fig. 2a)15,61. Moreover, it is also argued by some scholars that the motif of the lion on this carpet is from the Buddhist art of India, as lions are frequently mentioned in Buddhist stories and a Buddhist temple was also found at the site of Yingpan (Fig. 6)17,77.In addition, though the main decorative patterns on Yingpan Man’s coffin are the “Lianbi” pattern, images of vines, leaves, pomegranate flowers and vases were also depicted inside of the diamond patterns (Fig. S1)15. Among which, the image of pomegranate flowers is clearly an imported element from the West as pomegranates were originally domesticated in the middle East ~ 5000 years ago and were not introduced into China until the Han to Jin Dynasties, or with the flow of goods along the Silk Road78. In particular, the pomegranate was used as a symbol of health, fertility and rebirth as mentioned in many ancient cultures, especially in Greek and Turkish myths78. Thus, this is particularly interesting as the image of pomegranate appears not only on Yingpan Man’s coffin as flowers on the headboard and footboard (Fig. 6; Fig. S1), but also on his caftan as decorative patterns of fruited-trees (Figs. 6, 9). Specifically, the decorative patterns on Yingpan Man’s woolen caftan consists of six sets of nude puttis and animals with fruited pomegranate trees standing in between (Fig. 9)14. In particular, each of these six sets of images is composed of a symmetrical pair of confronting muscular puttis or animals (goats or bulls) that are either leaning away from or toward each other14. The nude puttis are holding either a spear, sword or a shield with capes swirling from their shoulders, while the animals of goats and bulls are in the pose of jumping and the bulls have laurel wreaths around their waists (Fig. 9)14. A similar design of stance and composition of figures was also discovered in a mosaic floor from the Villa of Good Fortune at Olynthos, Greece (paired female figures with weapons, fourth century BC)79 as well as another mosaic floor from Pella, the Macedonian capital (two nude youths with capes flying on their shoulder and weapons held in their hand, about to attack an animal in between of them, 325 to 300 BC)79. Thus, it is suggested that this caftan was the work of a weaver familiar with both Western and Eastern motifs as the character and poses of the nude puttis are clearly Western in style and appearance, the fruited pomegranate tree is believed to be a Persian motif, while the paired facing animals of goats and bulls are similar to the animal art of Central Asia8. However, this caftan was completed using the technique of double-weaving, the lining and belt were made of silk14, while the design of the side slits are indicative of a localized adaptation for horse-riding15. Moreover, analysis using high performance liquid chromatography (HPLC) on the red and yellow threads of this caftan suggest that they were dyed in a local workshop with indigenous materials of Rubia Tinctorum and Populus Pruinosa Schrenk, respectively80,81. In conclusion, the design and style of Yingpan Man’s caftan is unprecedented, and it is a masterpiece that combines both Greek/Roman, Persian, Central Asian, Chinese and local elements (Fig. 6).Figure 9Photo and sketch of the decorative patterns on Yingpan Man’s caftan. (The original pictures and sketches were previously published14 and provided by Wenying Li. Pictures were modified using Adobe Photoshop CC 2015 V.1.2. The final layout was created in Adobe Illustrator CC 2019 V.23.1.1).Full size imageThe opulence and fine quality of the objects buried with Yingpan Man indicate that he must have had a high social status before death14,47. Given the importance of the town of Yingpan as a trading center on the Silk Road, the excavators who discovered Yingpan Man suggested that he was a wealthy merchant from the West14. Others have proposed that Yingpan Man might have been a Sogdian merchant since the Sogdians (an Iranian-speaking people whose homeland lay near Samarkand in what is now Uzbekistan) were the richest traders along the route7. However, given the relatively young age of Yingpan Man before his death (~ 30 years old), it is unlikely that he amassed all his fortune and high social status only through trade, and possibly by inheritance or military feats. The government of the Jin Dynasty established the administrative organization of “Xiyu Zhangshi Fu” (meaning “Chief Governor of the Western Regions”) in ancient Xinjiang, and the capital city of “Xiyu Zhangshi Fu” was located nearby Lop Nur and is very close to the ancient city of Yingpan (~ 185 km away)15,16. In addition, comparison of Yingpan Man’s burial to other contemporary burials from Gansu also suggest that Yingpan Man was possibly a military official from the government of Central China16. More supporting evidence comes from the embroidered armband that was buried with Yingpan Man as colorful armbands were suggested to be used by soldiers for the protection from evil forces in ancient Xinjiang (Fig. 6)15. An additional explanation is that Yingpan Man was a noble or even a king of the nearby state named Shan (a.k.a. Moshan)76, and the ancient city of Yingpan was suggested to be the capital city of this state15. An alternative explanation is that Yingpan Man belonged to a local noble family who were displaced from Bactria to the southern Tarim Basin after civil strife in the Kushan Empire at the end of the second century AD, given the popularity of Kushan arts in this area during the Han to Jin Dynasties82,83. The isotopic evidence presented here, in particular the hair δ34S results, add additional information to Yingpan Man’s identity. The lack of δ34S isotopic variability (10.4‰ to 11.9‰) over the last ~ 3–4 years of life indicates that Yingpan Man was not a Silk Road traveler or merchant, at least during this period of his life. Thus, Yingpan Man appears to have been a local, possibly a governmental official or royal to this region of the Tarim Basin, perhaps from the nearby state of Shan. This might suggest why he was buried in the Yingpan cemetery as it was purported to be capital of this ancient state. More

The prevalence of S. aureus isolation from the different examined samplesStaphylococcal infections represent a public health issue in hospitals and health care settings as well as a major economical and welfare problem in dairy animal farming25. The prevalence of S. aureus isolation from the farm under the study (Table 2) showed that 63 (33.1%) out of 190 different samples were bacteriologically positive. Moreover, the isolation was mainly obtained from manager swabs followed by milk machine swabs, nasal swabs and hand swabs (60.0, 53.3, 40.0 and 28.0%, respectively), and to a lesser extent in milk samples (24.0%). Meanwhile it was not isolated at any percent from water trough swabs, at X2 = 48.8, P  More

Generation timeDuring the infectious period, an infected individual may produce a secondary infection. However, the individual’s infectiousness is not constant during the infectious period, but it can be approximated by the probability distribution of the generation time (GT), which accounts for the time between the infection of a primary case and the infection of a secondary case. Unfortunately, such distribution is not as easy to estimate as that of the serial interval, which accounts for the time between the onset of symptoms in a primary case to the onset of symptoms of a secondary case. This is because the time of infection is more difficult to detect than the time of symptoms onset. Ganyani et al.27 developed a methodology to estimate the distribution of the GT from the distributions of the incubation period and the serial interval. Assuming an incubation period following a gamma distribution with a mean of 5.2 days and a standard deviation (SD) of 2.8 days, they estimated the serial interval from 91 and 135 pairs of documented infector-infectee in Singapore and Tianjin (China). Then, they found that the GT followed a gamma distribution with mean = 5.20 (95% CI = [3.78, 6.78]) days and SD = 1.72 (95% CI = [0.91, 3.93]) for Singapore (hereafter GT1), and with mean = 3.95 (95% CI = [3.01, 4.91]) days and SD = 1.51 (95% CI = [0.74, 2.97]) for Tianjin (hereafter GT2). Ng et al.28 applied the same methodology to 209 pairs of infector-infectee in Singapore and determined a gamma distribution with mean = 3.44 (95% CI = [2.79, 4.11]) days and SD 2.39 (95% CI = [1.27, 3.45]; hereafter GT3). Figure 3 shows the probability density functions (PDF) of such distributions, fGT. The differences between them are remarkable. For example, the 54.5%, 81.0%, and 80.7% of the contagions are produced in a pre-symptomatic stage (in the first 5.2 days after primary infection) assuming GT1, GT2, and GT3, respectively.Figure 3Probability density function of the generation time distribution, fGT(t), of GT1 (blue line; Singapore27), GT2 (yellow line; Tianjin27), GT3 (red line; Singapore28), and GTth (black line; theoretical distribution). Bars are the discretized version, (widetilde{{f_{GT} }}left( n right)), of the PDF of GTth.Full size imageTheoretically, assuming that the incubation periods of two individuals are independent and identically distributed, which is quite plausible, the expected/mean values of the GT and the serial interval should be equal29,30. The mean of the serial interval is easier to estimate than that of the GT. For that reason, we assume a mean serial interval as estimated from a meta-analysis of 13 studies involving a total of 964 pairs of infector-infectee, which is 4.99 days (95% CI = [4.17, 5.82])31, is more reliable than the aforementioned means of the GT. This value is within the error estimates of the means of GT1 and GT2, but not for GT3. Then, we construct a theoretical distribution for the GT that follows a gamma distribution (hereafter GTth) with mean = 4.99 days and SD = 1.88 days. This theoretical distribution can be seen in Fig. 3 and approximates the average PDF of three gamma distributions with mean = 4.99 and the SD of GT1, GT2, and GT3. We assume a conservative CI = [1.51, 2.39] for the theoretical SD, defined with the minimum and maximum SD values of GT1, GT2, and GT3. GTth shows 63.1% of pre-symptomatic contagions.
R

0

from r
In theory, the basic reproduction number R0 can be estimated as far as the intrinsic growth rate r, and the distributions of both the latent and infectious periods are known26,32,33,34. The latent period accounts for the period during which an infected individual cannot infect other individuals. It is observed in diseases for which the infectious period starts around the end of the incubation period, as happened with influenza35 and SARS36. However, from Fig. 3 it is inferred that COVID-19 is transmissible from the moment of infection, and we will assume a null latent period. Then, if the GT follows a gamma distribution, R0 can be estimated from the formulation of Anderson and Watson32, which was adapted to null latent periods by Yan26 as$$ R_{0} = frac{{mean_{GT} }}{{1 – left( {1 + mean_{GT} cdot r cdot frac{1}{{shape_{GT} }}} right)^{{ – shape_{GT} }} }} cdot r, $$
(4)
where meanGT is the mean GT and shapeGT is one of the two parameters defining the gamma distribution, which can be estimated as$$ shape_{GT} = frac{{left( {mean_{GT} } right)^{2} }}{{left( {SD_{GT} } right)^{2} }}. $$
(5)
For GTth, we get R0 = 1.50 (CI = [1.41, 1.61]) for REMEDID I(n) and R0 = 1.76 (CI = [1.60, 1.94]) for official I(n). For the other three GT distributions, R0 ranges from 1.39 (CI = [1.27, 1.58]) to 1.51 (CI = [1.34, 1.80]) for REMEDID I(n) and from 1.59 (CI = [1.40, 1.88]) to 1.78 (CI = [1.51, 2.23]) for official I(n) (Table 1). In all cases, R0 from GTth are within those from the three known GT distributions and indistinguishable from them within the error estimates. The lower (upper) bound of the CI is estimated as the minimum (maximum) R0 obtained from all the possible combinations of 100 evenly spaced values covering the CI of r, meanGT and SDGT. Then, following the Bonferroni correction, the reported CI present at least a 85% of confidence level for GT1, GT2, and GT3, but it cannot be assured for GTth since the CI of its SD is unknown. In general, all these R0 estimates are lower than those summarised by Park et al.20.Table 1 R0 and HIT values of the ancestral SARS-CoV-2 variant estimated from GT1, GT2, GT3, and GTth, and REMEDID and official infections. For date0, “Dec.” means December 2019, and “Jan.” means January 2020.Full size tableAlternatively, R0 can be estimated by applying the Euler–Lotka equation29,33,$$ R_{0} = frac{1}{{mathop smallint nolimits_{0}^{ + infty } e^{ – rt} cdot f_{GT} left( t right)dt}}. $$
(6)
In this case, we get values closer to previous estimates20. In particular, for GTth, we get R0 = 2.12 (CI = [1.81, 2.48]) for REMEDID I(n) and R0 = 2.92 (CI = [2.28, 3.75]) for official I(n). For the other three GT distributions, R0 ranges from 1.63 (CI = [1.43, 1.90]) to 2.21 (CI = [1.59, 2.95]) for REMEDID I(n) and from 1.97 (CI = [1.59, 2.54]) to 3.11 (CI = [1.84, 4.90]) for official I(n) (Table 1). The CI are estimated as in Eq. (4).R0 from a dynamical modelWe designed a dynamic model with Susceptible-Infected-Recovered (SIR) as stocks that accounts for the infectiousness of the infectors. Such a model is a generalisation of the Susceptible-Exposed-Infected-Recovered (SEIR) model37. Births, deaths, immigration and emigration are ignored, which seems reasonable since the timescale of the outbreak is too short to produce significant demographic changes. For the sake of simplicity, the recovered stock includes recoveries and fatalities, and it is denoted as R(t). A random mixing population is assumed, that is a population where contacts between any two people are equally probable. Time is discretized in days, so the real time variable t is replaced by the integer variable n. As a consequence, the derivatives in the differential equations defining the dynamic model explained below are discrete derivatives.The size of the population is fixed at N = 100,000, and then, for any day n we get$$ tilde{S}left( n right) + left( {mathop sum limits_{k = 0}^{20} tilde{I}left( n-k right)} right) + tilde{R}left( n right) = N, $$
(7)
where (tilde{S}left( n right)), (tilde{I}left( n right)), and (tilde{R}left( n right)) are the discretized versions of S(t), I(t), and R(t) and (tilde{I}) is assumed to be null for negative integers. The summation is a consequence of the infectiousness, which is approximated according to the GT, whose PDF is discretized as$$ widetilde{{f_{GT} }}left( n right) = mathop smallint limits_{n – 1}^{n} f_{GT} left( t right) dt, $$
(8)
from n = 1 to 20. Figure 3 shows (widetilde{{f_{GT} }}left( n right)) for GTth. Truncating at n = 20 accounts for 99.99% of the area below the PDF of all the GT. Then, an infected individual at day n0 is expected to produce on average$$ widetilde{{R_{e} }}left( {n_{0} + n} right) cdot widetilde{{f_{GT} }}left( n right) $$
(9)
infections n days later, where (widetilde{{R_{e} }}left( n right)) is the discretized version of Re(t). From this expression, it is obvious that values of (widetilde{{R_{e} }}left( n right) < 1) will produce a decline of infections. Conversely, infections at day n0 are produced by all individuals infected during the previous 20 days as$$ tilde{I}(n_{0} ) = tilde{R}_{e} left( {n_{0} } right) cdot left( {mathop sum limits_{n = 1}^{20} tilde{I}left( {n_{0} - n} right) cdot widetilde{{f_{GT} }}left( n right)} right), $$ (10) whose continuous version has been reported in previous studies29,38. The expression in brackets is called total infectiousness of infected individuals at day n039. According to Eq. (1), Eq. (10) can be expressed in terms of R0 as$$ tilde{I}(n_{0} ) = R_{0} cdot frac{{tilde{S}left( {n_{0} } right)}}{N} cdot left( {mathop sum limits_{n = 1}^{20} tilde{I}left( {n_{0} - n} right) cdot widetilde{{f_{GT} }}left( n right)} right). $$ (11) As we want a dynamic model capable of providing (tilde{I}left( {n_{0} } right)) from the stocks at time step n0 − 1, we replaced (tilde{S}left( {n_{0} } right)) by (tilde{S}left( {n_{0} - 1} right)) in Eq. (11). This assumption makes sense in a discrete domain since the infections at time n0 take place in the susceptible population at time n0 − 1. Then, assuming that all stocks are set to zero for negative integers, our dynamic model can be expressed in terms of Eq. (7) and the following differential equations:$$ delta tilde{I}(n_{0} ) = R_{0} cdot frac{{tilde{S}left( {n_{0} - 1} right)}}{N} cdot left( {mathop sum limits_{n = 1}^{20} tilde{I}left( {n_{0} - n} right) cdot widetilde{{{text{f}}_{GT} }}left( n right)} right) - tilde{I}(n_{0} - 1), $$ (12) $$ delta tilde{S}left( {n_{0} } right) = {-}tilde{I}left( {n_{0} } right), $$ (13) $$ delta tilde{R}left( {n_{0} } right) = tilde{I}left( {n_{0} - 21} right), $$ (14) where (delta tilde{I}), (delta tilde{S}), and (delta tilde{R}) are the (discrete) derivatives of (tilde{I}), (tilde{S}), and (tilde{R}), respectively. Applying the initial conditions (tilde{S}left( 0 right) = N - 1), (tilde{I}left( 0 right) = 1), and (tilde{R}left( 0 right) = 0), it is assumed that the outbreak was produced by only one infector. The latter is not true in Spain, since several independent introductions of SARS-CoV-2 were detected40. However, for modelling purposes it is equivalent to introducing a single infection at day 0 or M infections produced by the single infection n days later. Then, the date of the initial time n = 0 is accounted as a parameter date0, which is optimised, as well as R0, to minimise the root-mean square of the residual between the model simulated (tilde{I}left( n right)) and the REMEDID and official I(n) for the period from date0 to n0.The model was implemented in Stella Architect software v2.1.1 (www.iseesystems.com) and exported to R software v4.1.1 with the help of deSolve (v1.28) and stats (v4.1.1) packages, and the Brent optimisation algorithm was implemented. For REMEDID I(n) and GTth, we obtained date0 = 13 December 2019 and R0 = 2.71 (CI = [2.33, 3.15]). Optimal solutions combine lower/higher R0 and earlier/later date0 (Fig. 4), which highlights the importance of providing an accurate first infection date to estimate R0. When the other three GT distributions were considered, we obtained similar date0, ranging from 12 to 17 December 2019, and R0 values ranging from 2.08 (CI = [1.86, 2.42]) to 2.85 (CI = [2.05, 3.25]; see Table 1). For official infections, date0 was set to 1 January 2020 for all cases, and R0 ranged from 1.81 (CI = [1.64, 2.07]) to 2.41 (CI = [1.80, 2.91]). The CI are estimated as in Eq. (4).Figure 4Root-mean square (RMS) of the residuals between infections from the model, which depends on date0 (x-axis) and R0 (y-axis), and REMEDID (from MoMo ED) and official infections. Parameters optimizing the model are highlighted in purple. RMS larger than 1275 (left panel) and 103 (right panel) are saturated in white.Full size image More

1.Federal Office for the Environment (FOEN). Action Plan for the Swiss Biodiversity Strategy (FOEN, Bern, 2017).2.Geldmann, J. & González-Varo, J. P. Conserving honey bees does not help wildlife. Science 359, 392–393 (2018).Article 

Google Scholar 
3.Egerer, M. & Kowarik, I. Confronting the modern gordian knot of urban beekeeping. Trends Ecol. Evol. 35, 956–959 (2020).Article 

Google Scholar 
4.Herrera, C. M. Gradual replacement of wild bees by honeybees in flowers of the Mediterranean Basin over the last 50 years. Proc. R. Soc. B Biol. Sci. 287, 16–20 (2020).
Google Scholar 
5.Torné-Noguera, A., Rodrigo, A., Osorio, S. & Bosch, J. Collateral effects of beekeeping: impacts on pollen-nectar resources and wild bee communities. Basic Appl. Ecol. 17, 199–209 (2016).Article 

Google Scholar 
6.Magrach, A., González-Varo, J. P., Boiffier, M., Vilà, M. & Bartomeus, I. Honeybee spillover reshuffles pollinator diets and affects plant reproductive success. Nat. Ecol. Evol. 1, 1299–1307 (2017).Article 

Google Scholar 
7.Prendergast, K. S., DIxon, K. W. & Bateman, P. W. Interactions between the introduced European honey bee and native bees in urban areas varies by year, habitat type and native bee guild. Biol. J. Linn. Soc. 133, 725–743 (2021).Article 

Google Scholar 
8.Herbertsson, L., Lindström, S. A. M., Rundlöf, M., Bommarco, R. & Smith, H. G. Competition between managed honeybees and wild bumblebees depends on landscape context. Basic Appl. Ecol. 17, 609–616 (2016).Article 

Google Scholar 
9.Stevenson, P. C. et al. The state of the world’s urban ecosystems: what can we learn from trees, fungi, and bees? Plants People Planet 2, 482–498 (2020).Article 

Google Scholar 
10.Ropars, L., Dajoz, I., Fontaine, C., Muratet, A. & Geslin, B. Wild pollinator activity negatively related to honey bee colony densities in urban context. PLoS ONE 14, e0222316 (2019).CAS 
Article 

Google Scholar 
11.Venter, Z. S. & Sydenham, M. A. K. Continental-scale land cover mapping at 10 m resolution over Europe (ELC10). Remote Sens. 13, 2301 (2021).Article 

Google Scholar 
12.Hardin, G. The tragedy of the commons. Science 162, 1243–1248 (1968).CAS 
Article 

Google Scholar 
13.Tew, N. E. et al. Quantifying nectar production by flowering plants in urban and rural landscapes. J. Ecol. 109, 1747–1757 (2021).Article 

Google Scholar 
14.Baldock, K. C. R. et al. A systems approach reveals urban pollinator hotspots and conservation opportunities. Nat. Ecol. Evol. 3, 363–373 (2019).Article 

Google Scholar 
15.Casanelles-Abella, J. et al. Applying predictive models to study the ecological properties of urban ecosystems: A case study in Zürich, Switzerland. Landsc. Urban Plan. 214, 104137 (2021).Article 

Google Scholar 
16.IPBES. Summary for policymakers of the global assessment report on biodiversity and ecosystem services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES, 2019).17.IUCN. IUCN’s Key Messages. First Draft of the Post-2020 Global Biodiversity Framework. In Convention on Biological Diversity Third meeting of the Open-Ended Working Group on the Post-2020 Global Biodiversity Framework (OEWG3) 15 (IUCN, 2021).18.Baldock, K. C. R. Opportunities and threats for pollinator conservation in global towns and cities. Curr. Opin. Insect Sci. 38, 63–71 (2020).Article 

Google Scholar 
19.Henry, M. & Rodet, G. The apiary influence range: a new paradigm for managing the cohabitation of honey bees and wild bee communities. Acta Oecologica 105, 103555 (2020).Article 

Google Scholar 
20.Ignatieva, M. & Hedblom, M. An alternative urban green carpet. Science 362, 148–149 (2018).CAS 
Article 

Google Scholar 
21.Vega, K. A. & Küffer, C. Promoting wildflower biodiversity in dense and green cities: The important role of small vegetation patches. Urban For. Urban Green. 62, 127165 (2021).Article 

Google Scholar 
22.Fabián, D., González, E., Sánchez Domínguez, M. V., Salvo, A. & Fenoglio, M. S. Towards the design of biodiverse green roofs in Argentina: assessing key elements for different functional groups of arthropods. Urban For. Urban Green. 61 (2021).23.Vega, K. A., Schläpfer-Miller, J. & Kueffer, C. Discovering the wild side of urban plants through public engagement. Plants People Planet 3, 389–401 (2021).Article 

Google Scholar 
24.Von Büren, R. S., Oehen, B., Kuhn, N. J. & Erler, S. High-resolution maps of Swiss apiaries and their applicability to study spatial distribution of bacterial honey bee brood diseases. PeerJ 2019, 1–21 (2019).
Google Scholar 
25.QGIS Development Team. QGIS Geographic Information System (Open Source Geospatial Foundation Project, 2020).26.R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, 2019).27.R Studio Team. R studio: Integrated Development for R (RStudio, Boston, 2020).28.Casanelles-Abella, J. & Moretti M. Challenging the sustainability of urban beekeeping: evidence from Swiss cities. Envidat. https://doi.org/10.16904/envidat.239 (2021).29.Casanelles-Abella, J. Code for the paper: Challenging the sustainability of urban beekeeping: evidence from Swiss cities (v. 1.0). Zenodo https://doi.org/10.5281/zenodo.5618254 (2021).30.Swiss Fedearl Office of Topography Swisstopo. SWISSIMAGE 25. https://www.swisstopo.admin.ch/en/geodata/images/ortho/swissimage25.html#links (Swisstopo, 2021). More

1.Lindow SE, Brandl MT. Microbiology of the phyllosphere. Appl Environ Microbiol. 2003;69:1875LP–1883.
Google Scholar 
2.Morella NM, Zhang X, Koskella B. Tomato seed-associated bacteria confer protection of seedlings against foliar disease caused by Pseudomonas syringae. Phytobiomes J. 2019;3:177–90.
Google Scholar 
3.Innerebner G, Knief C, Vorholt JA. Protection of Arabidopsis thaliana against leaf-pathogenic Pseudomonas syringae by Sphingomonas strains in a controlled model system. Appl Environ Microbiol. 2011;77:3202–10.CAS 
PubMed 
PubMed Central 

Google Scholar 
4.Fu S-F, Sun P-F, Lu H-Y, Wei J-Y, Xiao H-S, Fang W-T, et al. Plant growth-promoting traits of yeasts isolated from the phyllosphere and rhizosphere of Drosera spatulata Lab. Fungal Biol. 2016;120:433–48.CAS 
PubMed 

Google Scholar 
5.Laforest-Lapointe I, Paquette A, Messier C, Kembel SW. Leaf bacterial diversity mediates plant diversity and ecosystem function relationships. Nature. 2017;546:145–7.CAS 
PubMed 

Google Scholar 
6.Lindow SE, Leveau JHJ. Phyllosphere microbiology. Curr Opin Biotechnol. 2002;13:238–43.CAS 
PubMed 

Google Scholar 
7.Fürnkranz M, Wanek W, Richter A, Abell G, Rasche F, Sessitsch A. Nitrogen fixation by phyllosphere bacteria associated with higher plants and their colonizing epiphytes of a tropical lowland rainforest of Costa Rica. ISME J. 2008;2:561–70.PubMed 

Google Scholar 
8.Ottesen AR, Gorham S, Reed E, Newell MJ, Ramachandran P, Canida T, et al. Using a control to better understand phyllosphere microbiota. PLoS ONE. 2016;11:e0163482.PubMed 
PubMed Central 

Google Scholar 
9.Jones JDG, Dangl JL. The plant immune system. Nature. 2006;444:323–9.CAS 
PubMed 

Google Scholar 
10.Bodenhausen N, Bortfeld-miller M, Ackermann M, Vorholt JA. A synthetic community approach reveals plant genotypes affecting the phyllosphere microbiota. PLoS Biol. 2014; 10. https://doi.org/10.1371/journal.pgen.1004283.11.Horton MW, Bodenhausen N, Beilsmith K, Meng D, Muegge BD, Subramanian S, et al. Genome-wide association study of Arabidopsis thaliana leaf microbial community. Nat Commun. 2014;5:5320.PubMed 

Google Scholar 
12.Hacquard S, Spaepen S, Garrido-Oter R, Schulze-Lefert P. Interplay between innate immunity and the plant microbiota. Annu Rev Phytopathol. 2017;55:565–89.CAS 
PubMed 

Google Scholar 
13.Zhalnina K, Louie KB, Hao Z, Mansoori N, Nunes U, Shi S et al. Dynamic root exudate chemistry and substrate preferences drive patterns in rhizosphere microbial community assembly. Nat Microbiol. 2018. https://doi.org/10.1038/s41564-018-0129-3.14.Humphrey PT, Whiteman NK. Insect herbivory reshapes a native leaf microbiome. Nat Ecol Evol. 2020;4:221–9.PubMed 
PubMed Central 

Google Scholar 
15.Yadav RKP, Karamanoli K, Vokou D. Bacterial colonization of the phyllosphere of Mediterranean perennial species as influenced by leaf structural and chemical features. Micro Ecol. 2005;50:185–96.CAS 

Google Scholar 
16.Morella NM, Weng FCH, Joubert PM, Metcalf CJE, Lindow S, Koskella B. Successive passaging of a plant-associated microbiome reveals robust habitat and host genotype-dependent selection. Proc Natl Acad Sci USA. 2020;117:1148–59.CAS 
PubMed 

Google Scholar 
17.Wagner MR, Busby PE, Balint-Kurti P. Analysis of leaf microbiome composition of near-isogenic maize lines differing in broad-spectrum disease resistance. N Phytol. 2019;225:2152–65.
Google Scholar 
18.Wagner MR, Lundberg DS, Del Rio TG, Tringe SG, Dangl JL, Mitchell-Olds T. Host genotype and age shape the leaf and root microbiomes of a wild perennial plant. Nat Commun. 2016;7:1–15.CAS 

Google Scholar 
19.Horner-Devine MC, Bohannan BJM. Phylogenetic clustering and overdispersion in bacterial communities. Ecology. 2006;87:S100–8.PubMed 

Google Scholar 
20.Kembel SW, O’Connor TK, Arnold HK, Hubbell SP, Wright SJ, Green JL. Relationships between phyllosphere bacterial communities and plant functional traits in a neotropical forest. Proc Natl Acad Sci USA 2014; 1–6.21.Burns AR, Stephens WZ, Stagaman K, Wong S, Rawls JF, Guillemin K, et al. Contribution of neutral processes to the assembly of gut microbial communities in the zebrafish over host development. ISME J. 2016;10:655–64.CAS 
PubMed 

Google Scholar 
22.Sloan WT, Woodcock S, Lunn M, Head IM, Curtis TP. Modeling taxa-abundance distributions in microbial communities using environmental sequence data. Micro Ecol. 2007;53:443–55.
Google Scholar 
23.Laforest-Lapointe I, Messier C, Kembel SW. Host species identity, site and time drive temperate tree phyllosphere bacterial community structure. Microbiome 2016; 1–10.24.Schlaeppi K, Dombrowski N, Oter RG, Ver Loren van Themaat E, Schulze-Lefert P. Quantitative divergence of the bacterial root microbiota in Arabidopsis thaliana relatives. Proc Natl Acad Sci USA. 2014;111:585LP–592.
Google Scholar 
25.Gallart M, Adair KL, Love J, Meason DF, Clinton PW, Xue J, et al. Host genotype and nitrogen form shape the root microbiome of Pinus radiata. Micro Ecol. 2018;75:419–33.CAS 

Google Scholar 
26.Hambäck PA, Inouye BD, Andersson P, Underwood N. Effects of plant neighborhoods on plant–herbivore interactions: resource dilution and associational effects. Ecology. 2014;95:1370–83.PubMed 

Google Scholar 
27.Underwood N, Inouye BD, Hambäck PA. A conceptual framework for associational effects: when do neighbors matter and how would we know? Q Rev Biol. 2014;89:1–19.PubMed 

Google Scholar 
28.Barbosa P, Hines J, Kaplan I, Martinson H, Szczepaniec A, Szendrei Z. Associational resistance and associational susceptibility: having right or wrong neighbors. Annu Rev Ecol Evol Syst. 2009;40:1–20.
Google Scholar 
29.Janzen DH. Herbivores and the number of tree species in tropical forests. Am Nat. 1970;104:501–28.
Google Scholar 
30.Connell JH. On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. in Den Boer PJ, Gradwell G, editors. Dynamics of populations. PUDOC, 1971, p. 298–312.31.Mangan SA, Schnitzer SA, Herre EA, Mack KML, Valencia MC, Sanchez EI, et al. Negative plant–soil feedback predicts tree-species relative abundance in a tropical forest. Nature. 2010;466:752–5.CAS 
PubMed 

Google Scholar 
32.Miller EC, Perron GG, Collins CD. Plant‐driven changes in soil microbial communities influence seed germination through negative feedbacks. Ecol Evol. 2019;0:1–14.
Google Scholar 
33.Antonovics J, Ellstrand NC. Experimental studies of the evolutionary significance of sexual reproduction. I. A test of the frequency-dependent selection hypothesis. Evolution. 1984;38:103–15.PubMed 

Google Scholar 
34.Ellstrand NC, Antonovics J. Experimental studies of the evolutionary significance of sexual reproduction II. A test of the density-dependent selection hypothesis. Evolution. 1985;39:657–66.PubMed 

Google Scholar 
35.Naeem S, Tjossem SF, Byers D, Bristow C, Li S. Plant neighborhood diversity and production. Ecoscience. 1999;6:355–65.
Google Scholar 
36.Worrich A, Musat N. Associational effects in the microbial neighborhood. ISME J 2019; 2143–9.37.Copeland JK, Yuan L, Layeghifard M, Wang PW, Guttman DS. Seasonal community succession of the phyllosphere microbiome. Mol Plant Microbe Interact. 2015;28:274–85.CAS 
PubMed 

Google Scholar 
38.Lajoie G, Kembel SW. Host neighborhood shapes bacterial community assembly and specialization on tree species across a latitudinal gradient. Ecol Monogr. 2021;0:1–18.
Google Scholar 
39.Lymperopoulou D, Adams R, Lindow SE, Löffler F. Contribution of vegetation to the microbial composition of nearby outdoor air. Appl Environ Microbiol. 2016;82:3822–33.CAS 
PubMed 
PubMed Central 

Google Scholar 
40.Lindow SE, Andersen G. Influence of immigration on epiphytic bacterial populations on navel orange leaves. Appl Environ Microbiol. 1996;62:2978–87.CAS 
PubMed 
PubMed Central 

Google Scholar 
41.Massoni J, Bortfeld-miller M, Widmer A, Vorholt JA. Capacity of soil bacteria to reach the phyllosphere and convergence of floral communities despite soil microbiota variation. Proc Natl Acad Sci USA 2021; 118. https://doi.org/10.1073/pnas.2100150118.42.Leibold MA, Holyoak M, Mouquet N, Amarasekare P, Chase JM, Hoopes MF, et al. The metacommunity concept: a framework for multi-scale community ecology. Ecol Lett. 2004;7:601–13.
Google Scholar 
43.Fodelianakis S, Lorz A, Valenzuela-cuevas A, Barozzi A, Booth JM, Daffonchio D. Dispersal homogenizes communities via immigration even at low rates in a simplified synthetic bacterial metacommunity. Nat Commun. 2019;10:1–12.
Google Scholar 
44.Burns AR, Miller E, Agarwal M, Rolig AS, Milligan-Myhre K, Seredick S et al. Interhost dispersal alters microbiome assembly and can overwhelm host innate immunity in an experimental zebrafish model. Proc Natl Acad Sci USA 2017;114. https://doi.org/10.1073/pnas.1702511114.45.Chelius MK, Triplett EW. The diversity of archaea and bacteria in association with the roots of Zea mays L. Micro Ecol. 2001;41:252–63.CAS 

Google Scholar 
46.Bodenhausen N, Horton MW, Bergelson J. Bacterial communities associated with the leaves and the roots of Arabidopsis thaliana. PLoS ONE. 2013;8:e56329.CAS 
PubMed 
PubMed Central 

Google Scholar 
47.Lundberg DS, Yourstone S, Mieczkowski P, Jones CD, Dangl JL. Practical innovations for high-throughput amplicon sequencing. Nat Methods. 2013;10:999–1002.CAS 
PubMed 

Google Scholar 
48.Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: high-resolution sample inference from Illumina amplicon data. Nat Methods. 2016;13:581.CAS 
PubMed 
PubMed Central 

Google Scholar 
49.R Core Team. R: A language and environment for statistical computing. 2020. http://cran.r-project.org.50.Morgan M, Anders S, Lawrence M, Aboyoun P, Pagès H, Gentleman R. ShortRead: a bioconductor package for input, quality assessment and exploration of high-throughput sequence data. Bioinformatics. 2009;25:2607–8.CAS 
PubMed 
PubMed Central 

Google Scholar 
51.Pagès H, Aboyoun P, Gentleman R, DebRoy S. Biostrings: efficient manipulation of biological strings. 2020.52.McMurdie P, Holmes S. phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE 2013; 8. https://doi.org/10.1371/journal.pone.0061217tle.53.Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol. 2007;73:5261–7.CAS 
PubMed 
PubMed Central 

Google Scholar 
54.Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41:D590–D596.CAS 
PubMed 

Google Scholar 
55.Davis NM, Proctor DM, Holmes SP, Relman DA, Callahan BJ. Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. Microbiome. 2018;6:226.PubMed 
PubMed Central 

Google Scholar 
56.Morella NM, Gomez AL, Wang G, Leung MS, Koskella B. The impact of bacteriophages on phyllosphere bacterial abundance and composition. Mol Ecol. 2018;27:2025–38.PubMed 

Google Scholar 
57.Oksanen J, Blanchet FG, Roeland K, Legendre P, Minchin P, O’Hara RB et al. vegan: Community ecology package. 2015. http://cran.r-project.org.58.Anderson MJ. A new method for non-parametric multivariate analysis of variance. Austral Ecol. 2001;26:32–46.
Google Scholar 
59.De Cáceres M, Legendre P. Associations between species and groups of sites: indices and statistical inference. Ecology. 2009;90:3566–74.PubMed 

Google Scholar 
60.Ersts PJ. Geographic Distance Matrix Generator. http://biodiversityinformatics.amnh.org/open_source/gdmg.61.Sprockett D. reltools: Microbiome Amplicon Analysis and Visualization. 2021.62.Sloan WT, Lunn M, Woodcock S, Head IM, Nee S, Curtis TP. Quantifying the roles of immigration and chance in shaping prokaryote community structure. Environ Microbiol. 2006;8:732–40.PubMed 

Google Scholar 
63.Wright ES. Using DECIPHER v2.0 to analyze big biological sequence data in R. R J. 2016;8:352–9.
Google Scholar 
64.Schliep KP. phangorn: phylogenetic analysis in R. Bioinformatics. 2011;27:592–3.CAS 
PubMed 

Google Scholar 
65.Kembel SW, Cowan PD, Helmus MR, Cornwell WK, Morlon H, Ackerly DD, et al. Picante: R tools for integrating phylogenies and ecology. Bioinformatics. 2010;26:1463–4.CAS 
PubMed 

Google Scholar 
66.Koskella B. The phyllosphere. Curr Biol. 2020;30:R1143–R1146.CAS 
PubMed 

Google Scholar 
67.Chaparro JM, Badri DV, Vivanco JM. Rhizosphere microbiome assemblage is affected by plant development. ISME J. 2014;8:790–803.CAS 
PubMed 

Google Scholar 
68.İnceoğlu Ö, Al-Soud WA, Salles JF, Semenov AV, van Elsas JD. Comparative analysis of bacterial communities in a potato field as determined by pyrosequencing. PLoS ONE. 2011;6:e23321.PubMed 
PubMed Central 

Google Scholar 
69.Christian N, Herre EA, Mejia LC, Clay K. Exposure to the leaf litter microbiome of healthy adults protects seedlings from pathogen damage. Proc R Soc B Biol Sci. 2017;284:20170641.
Google Scholar 
70.Leigh EG, Davidar P, Dick CW, Terborgh J, Puyravaud J-P, ter Steege H, et al. Why do some tropical forests have so many species of trees? Biotropica. 2004;36:447–73.
Google Scholar 
71.Hyatt LA, Rosenberg MS, Howard TG, Bole G, Fang W, Anastasia J, et al. The distance dependence prediction of the Janzen-Connell hypothesis: a meta-analysis. Oikos. 2003;103:590–602.
Google Scholar 
72.Carson W, Anderson J, Leigh E, Schnitzer S. Challenges associated with testing and falsifying the Janzen_Connell hypothesis: a review and critique. In: Carson W, Schnitzer SA, editors. Tropical forest community ecology. Wiley Blackwell; 2008. p. 210–41. More

Genome sequencing and assemblyGenome assemblies ranged in size from 799.9 Mbp in Melanocharis versteri to 1053.5 Mbp in Sericornis nouhuysi. The number of scaffolds ranged from 14,086 scaffolds in Melipotes ater to 87,957 scaffolds in Ficedula hyperythra and N50 ranged between ca. 40 Kbp to and 25 Mbp. Benchmarking Universal Single-Copy Orthologs (BUSCO) analyses of genome completeness ranged from a high proportion of complete BUSCOs in Melipotes ater, 86.8% to only 66.7% complete BUSCOs in Rhipidura albolimbata. For most species, the proportions of complete BUSCOs were 75–80%. Overall, the proportion of missing BUSCOs was low, ranging from 6.6% in Melipotes ater to 15.2% in Rhipidura albolimbata (see Supplementary Table 1 for all genome assembly statistics and Supplementary Fig. 1 for the number of SNP variants per species).Kinship analyses of individuals within populationsSampling of closely related individuals can dramatically bias estimates of population structure and demographics. Two Pachycephala schlegelii individuals (A117 and A118) showed a pairwise kinship coefficient of 0.144, indicative of being half-siblings. The two individuals were collected at the same locality on the same date. Similarly, two Ifrita kowaldi individuals (D116 and D117) showed a pairwise kinship coefficient of 0.135, also suggestive of being half-siblings. In this case, the individuals were collected on the same sampling locality on two consecutive days. To not bias downstream demographic analyses, one of the P. schlegelii (A118) and one of the I. kowaldi (D117) individuals were excluded from all subsequent analyses. For all other species, no closely related individuals were identified.Genetic differentiationEstimated levels of differentiation between populations were initially based on three approaches; (i) calculation of FST (the fixation index), which quantifies the degree of genetic differentiation between populations based on the variation in allele frequencies, ranging between 0 (complete mixing of individuals) and 1 (complete separation of populations) (Fig. 1), (ii) Standardized pairwise FST used to conduct a Principal Component Analysis (PCA) in order to visualize population structure (Supplementary Fig. 1) and (iii) Admixture analysis as implemented in STRUCTURE (a clustering algorithm that infers the most likely number of groups [K]), in which individuals are grouped into clusters according to the proportion of their ancestry components (Supplementary Fig. 1). As a preliminary analysis, we calculated FST and constructed PCA plots for the four congeneric (incl. Sericornis/Aethomyias [until recently placed in the genus Sericornis]) species pairs in our study (Supplementary Fig. 2), which were aligned using the same reference genome. This was done to ascertain that no samples had been misidentified and to gauge levels of differentiation between distinct species. All species were genetically well separated and FST values ranged from 0.08 for the two Ptiloprora species to 0.20 for the two Ficedula species.For five out of six species from Buru/Seram, genetic differentiation (FST) was high between islands (Fig. 1), and comparable to differentiation between named congeneric species in this study (e.g. Ptiloprora and Melipotes); Ceyx lepidus (FST = 0.16), Thapsinillas affinis (FST = 0.15), Ficedula buruensis (FST = 0.13) and Pachycephala macrorhyncha (FST = 0.09). In contrast, differentiation in Ficedula hyperythra was consistent with population-level differentiation (FST = 0.04). In all cases, individuals from Buru and Seram were clearly differentiated in the PCA and STRUCTURE plots (Supplementary Fig. 1A). For Ceyx lepidus, Ficedula buruensis and Pachycephala macrorhyncha, samples were collected at multiple elevations and we therefore calculated genetic differentiation between elevations (Buru: 1097 m versus 1435 m and Seram: 1000 m versus 1300 m) to determine any potential parapatric differentiation along the gradients. In all possible comparisons, FST values did not differ significantly from 0. Moreover, PCA plots showed that samples did not cluster according to elevation (Supplementary Fig. 3A).Three of the thirteen New Guinean population pairs occurring in Mount Wilhelm and Huon showed relatively high genetic divergences: Melipotes fumigatus/ater (FST = 0.08), Paramythia montium (FST = 0.09) and Ifrita kowaldi (FST = 0.07) (Fig. 1) with populations clearly separated (Supplementary Fig. 1). By contrast, the two lowland species Toxorhamphus novaeguineae and Melilestes megarhynchus showed little genetic differentiation, FST = 0.00. For the remaining species, genetic differentiation between Mount Wilhelm and Huon ranged between FST = 0.01–0.05. Despite this moderate level of genetic differentiation, the populations of Mount Wilhelm and Huon could be clearly distinguished in the PCA plots. In all cases STRUCTURE suggested a scenario with K = 2 with some mixing of individuals, except for Rhipidura albolimbata, in which K = 1 was suggested.For five bird species we included an additional population from Mount Scratchley, which is also situated in the Central Range but ~400 km to the southeast of Mount Wilhelm. Genetic differentiation of this population from the other two populations was comparable with that between Mount Wilhelm and Huon. The highest genetic differentiation was found in Paramythia montium (FST = 0.10 both between Mount Wilhelm and Mount Scratchley and between Huon and Mount Scratchley). In the case of Peneothello sigillata, the Mount Scratchley population appeared genetically well-differentiated from both the populations of Mount Wilhelm (FST = 0.06) and Huon (FST = 0.07). In both cases, STRUCTURE suggested a scenario of K = 3, with individual assignments matching the three geographically circumscribed populations. For Pachycephala schlegelii, genetic differentiation was relatively high between Huon and Mount Scratchley (FST = 0.05), but low between Mount Wilhelm and Mount Scratchley (FST = 0.01). Accordingly, STRUCTURE suggested a scenario with K = 2 groups. For the remaining two species Sericornis nouhuysi showed some differentiation (FST = 0.03) between Mount Wilhelm and Huon and Aethomyias papuensis showed minor differentiation (FST = 0.02 between Mount Scratchley and Huon (Supplementary Table 2), but for both species, STRUCTURE suggested a scenario of K = 2 with considerable mixing of individuals between populations.Samples from Mount Wilhelm were collected at elevations ranging from 1700 to 3700 m, again allowing us to test for differences within populations on a single slope, a finding that would be consistent with incipient parapatric speciation. No species showed significant differences in FST when comparing individuals from different elevations, and concordantly there was little clustering of individuals by elevation in the PCA plots. Even when individuals were collected as far as 2000 elevational meters apart (as in the case of Origma robusta), genetic differentiation was low (FST = 0.01). In Huon, all samples were collected at the same elevation, except for Ifrita kowaldi, for which genetic differentiation of FST = 0.03 was found between individuals collected at 2300 m and 2950 m (Supplementary Fig. 3B, Supplementary Table 2). These analyses however, suffer from very small sample sizes that hinder a thorough analysis of parapatric speciation events. Furthermore, we note that divergence with gene flow may not manifest as a genome-wide phenomenon (at least, not until the taxa are so differentiated that gene flow has ceased). Instead, it may proceed via selection acting to create small ‘islands of differentiation’ within the genome against a background of negligible differentiation22,23. Such analyses require large sample sizes and are therefore not possible herein.Correlations between genetic divergence and elevationIf lineages colonize mountains from the lowlands, followed by range contraction and differentiation in the highlands, we would expect a signature of larger genetic differentiation (FST) between populations inhabiting higher elevations. We found no relationship between genetic differentiation (FST) and the altitudinal floor (the lowest elevation at which a species/population occurs) for the five Moluccan species, but for all New Guinean taxa with the exception of Melipotes fumigatus/ater we found a significant positive correlation (r = 0.83, p  More

1.Kawecki, T. J. & Ebert, D. Conceptual issues in local adaptation. Ecol. Lett. 7, 1225–1241 (2004).
Google Scholar 
2.Fisher, T. W. et al. Handbook of Biological Control: Principles and Applications of Biological Control (Academic Press, London, 1999).
Google Scholar 
3.Dyck, V. A., Hendrichs, J. & Robinson, A. S. Sterile insect technique: Principles and practice in area-wide integrated pest management. In Sterile Insect Technique: Principles and Practice in Area-Wide Integrated Pest Management. https://doi.org/10.1007/1-4020-4051-2. (2005)4.Etges, W. J. & Noor, M. A. F. Genetics of Mate Choice: From Sexual Selection to Sexual Isolation. (Kluwer Academic Publishers, 2002).5.Harbach, R. E. Review of the internal classification of the genus Anopheles (Diptera: Culicidae): The foundation for comparative systematics and phylogenetic research. Bull. Entomol. Res. 84, 331–342 (1994).
Google Scholar 
6.Rogers, D. J., Randolph, S. E., Snow, R. W. & Hay, S. I. Satellite imagery in the study and forecast of malaria. Nature 415, 710–715 (2002).CAS 
PubMed 
PubMed Central 

Google Scholar 
7.Bayoh, M. N. N., Thomas, C. J. J. & Lindsay, S. W. W. Mapping distributions of chromosomal forms of Anopheles gambiae in West Africa using climate data. Med. Vet. Entomol. 15, 267–274 (2001).CAS 
PubMed 

Google Scholar 
8.Namountougou, M. et al. Multiple insecticide resistance in Anopheles gambiae s. l. Populations from Burkina Faso. West Africa. PLoS One 7, e48412 (2012).CAS 
PubMed 
ADS 

Google Scholar 
9.Benedict, M. Q. & Robinson, A. S. The first releases of transgenic mosquitoes: An argument for the sterile insect technique. Trends Parasitol. 19, 349–355 (2003).PubMed 

Google Scholar 
10.Maïga, H. et al. Mating competitiveness of sterile male Anopheles coluzzii in large cages. Malar. J. 13, 460 (2014).PubMed 
PubMed Central 

Google Scholar 
11.Clements, A. N. The Biology of Mosquitoes. Sensory Reception and Behaviour Behaviour, Vol. 2. (Wallingford, 1999).12.Doug, P. et al. Genetic and environmental factors associated with laboratory rearing affect survival and assortative mating but not overall mating success in Anopheles gambiae Sensu Stricto. PLoS One 8, e82631 (2013).
Google Scholar 
13.Baeshen, R. et al. Differential effects of inbreeding and selection on male reproductive phenotype associated with the colonization and laboratory maintenance of Anopheles gambiae. Malar. J. 13, 19 (2014).PubMed 
PubMed Central 

Google Scholar 
14.Bargielowski, I., Kaufmann, C., Alphey, L., Reiter, P. & Koella, J. Flight performance and teneral energy reserves of two genetically-modified and one wild-type strain of the yellow fever mosquito Aedes aegypti. Vector-Borne Zoonotic Dis. 12, 1053–1058 (2012).PubMed 
PubMed Central 

Google Scholar 
15.Harris, A. F. et al. Field performance of engineered male mosquitoes. Nat. Biotechnol. 29, 1034–1037 (2011).CAS 
PubMed 

Google Scholar 
16.Alphey, L. et al. Genetic control of Aedes mosquitoes. Pathogens and Global Health 107, 170–179 (2013).PubMed 
PubMed Central 

Google Scholar 
17.Lehmann, T. et al. Tracing the origin of the early wet-season Anopheles coluzzii in the Sahel. Evol. Appl. 10, 704–717 (2017).CAS 
PubMed 
PubMed Central 

Google Scholar 
18.Huestis, D. L. et al. Windborne long-distance migration of malaria mosquitoes in the Sahel. Nature 574, 404–408 (2019).CAS 
PubMed 

Google Scholar 
19.Wondji, C., Simard, F. & Fontenille, D. Evidence for genetic differentiation between the molecular forms M and S within the Forest chromosomal form of Anopheles gambiae in an area of sympatry. Insect Mol. Biol. 11, 11–19 (2002).
CAS 
PubMed 

Google Scholar 
20.Simard, F., Nchoutpouen, E., Toto, J. C. & Fontenille, D. Geographic distribution and breeding site preference of Aedes albopictus and Aedes aegypti (Diptera: Culicidae) in Cameroon, Central Africa. J. Med. Entomol. 42, 726–731 (2005).PubMed 

Google Scholar 
21.Roux, O., Diabaté, A. & Simard, F. Divergence in threat sensitivity among aquatic larvae of cryptic mosquito species. J. Anim. Ecol. 83, 702–711 (2014).PubMed 

Google Scholar 
22.Costantini, C. et al. Living at the edge: Biogeographic patterns of habitat segregation conform to speciation by niche expansion in Anopheles gambiae. BMC Ecol. 9 (2009).23.The Anopheles gambiae 1000 Genomes Consortium. Genetic diversity of the African malaria vector Anopheles gambiae. Nature 552, 96–100 (2017).PubMed Central 

Google Scholar 
24.Oliva, C. F., Benedict, M. Q., Lempérière, G. & Gilles, J. Laboratory selection for an accelerated mosquito sexual development rate. Malar. J. 10, 135 (2011).PubMed 
PubMed Central 

Google Scholar 
25.Munhenga, G. et al. Evaluating the potential of the sterile insect technique for malaria control: Relative fitness and mating compatibility between laboratory colonized and a wild population of Anopheles arabiensis from the Kruger National Park, South Africa. Parasit. Vectors 4, 208 (2011).PubMed 
PubMed Central 

Google Scholar 
26.Lee, H. L. et al. Mating compatibility and competitiveness of transgenic and wild type Aedes aegypti (L.) under contained semi-field conditions. Transgenic Res. 22, 47–57 (2013).CAS 
PubMed 

Google Scholar 
27.Damiens, D. et al. Cross-Mating compatibility and competitiveness among Aedes albopictus strains from distinct geographic origins-implications for future application of sit programs in the south west Indian ocean islands. PLoS One 11, e0163788 (2016).PubMed 
PubMed Central 

Google Scholar 
28.Zheng, X. et al. Incompatible and sterile insect techniques combined eliminate mosquitoes. Nature 572, 56–61 (2019).CAS 
ADS 

Google Scholar 
29.Aguilar, R. et al. Genome-wide analysis of transcriptomic divergence between laboratory colony and field Anopheles gambiae mosquitoes of the M and S molecular forms. Insect Mol. Biol. 19, 695–705 (2010).CAS 
PubMed 
PubMed Central 

Google Scholar 
30.Sawadogo, P. S. et al. Swarming behaviour in natural populations of Anopheles gambiae and An. coluzzii: Review of 4 years survey in rural areas of sympatry, Burkina Faso (West Africa). Acta Trop. 130, 24–34 (2014).
Google Scholar 
31.Poda, S. B. et al. Sex aggregation and species segregation cues in swarming mosquitoes: Role of ground visual markers. Parasit. Vectors 12, 589 (2019).PubMed 
PubMed Central 

Google Scholar 
32.Ekechukwu, N. E. et al. Heterosis increases fertility, fecundity, and survival of laboratory-produced F1 hybrid males of the malaria mosquito Anopheles coluzzii. G3 Genes Genomes Genet. 5, 2693–2709 (2015).CAS 

Google Scholar 
33.Ng’habi, K. R. et al. Colonization of malaria vectors under semi-field conditions as a strategy for maintaining genetic and phenotypic similarity with wild populations. Malar. J. 14, 10 (2015).PubMed 
PubMed Central 

Google Scholar 
34.Huho, B. J. et al. Nature beats nurture: A case study of the physiological fitness of free-living and laboratory-reared male Anopheles gambiae s.l. J. Exp. Biol. 210, 2939–2947 (2007).CAS 
PubMed 

Google Scholar 
35.Ferguson, H. M., John, B., Ng, K. & Knols, B. G. J. Redressing the sex imbalance in knowledge of vector biology. Trends Ecol. Evol. 20, 202–209 (2005).PubMed 

Google Scholar 
36.Hassan, M., El-Motasim, W. M., Ahmed, R. T. & El-Sayed, B. B. Prolonged colonisation, irradiation, and transportation do not impede mating vigour and competitiveness of male Anopheles arabiensis mosquitoes under semi-field conditions in Northern Sudan. Malar. World J. 1 (2010).37.Yamada, H., Vreysen, M. J. B., Gilles, J. R. L., Munhenga, G. & Damiens, D. D. The effects of genetic manipulation, dieldrin treatment and irradiation on the mating competitiveness of male Anopheles arabiensis in field cages. Malar. J. 13, 1–10 (2014).
Google Scholar 
38.Munhenga, G. et al. Mating competitiveness of sterile genetic sexing strain males (GAMA) under laboratory and semi-field conditions : Steps towards the use of the Sterile Insect Technique to control the major malaria vector Anopheles arabiensis in South Africa. Parasit. Vectors 9, 1–12 (2016).
Google Scholar 
39.Assogba, B. S. et al. Characterization of swarming and mating behaviour between Anopheles coluzzii and Anopheles melas in a sympatry area of Benin. Acta Trop. 132S, 1–11 (2013).
Google Scholar 
40.Charlwood, J. D. et al. The swarming and mating behaviour of Anopheles gambiae s.s. (Diptera: Culicidae) from São Tomé Island. J. Vector Ecol. 27, 178–183 (2002).CAS 
PubMed 

Google Scholar 
41.Diabate, A. et al. Natural swarming behaviour of the molecular M form of Anopheles gambiae. Trans. R. Soc. Trop. Med. Hyg. 97, 713–716 (2003).CAS 
PubMed 

Google Scholar 
42.Manoukis, N. C. et al. Structure and dynamics of male swarms of Anopheles gambiae. J. Med. Entomol. 46, 227–235 (2009).PubMed 

Google Scholar 
43.Aldersley, A. et al. Too ‘sexy’ for the field? Paired measures of laboratory and semi-field performance highlight variability in the apparent mating fitness of Aedes aegypti transgenic strains. Parasit. Vectors 12, 357 (2019).PubMed 
PubMed Central 

Google Scholar 
44.Pantoja-Sánchez, H., Gomez, S., Velez, V., Avila, F. W. & Alfonso-Parra, C. Precopulatory acoustic interactions of the New World malaria vector Anopheles albimanus (Diptera: Culicidae). Parasit. Vectors 12, 386 (2019).PubMed 
PubMed Central 

Google Scholar 
45.Gibson, G., Warren, B. & Russell, I. J. Humming in tune: Sex and species recognition by mosquitoes on the wing. JARO 540, 527–540 (2010).
Google Scholar 
46.Pennetier, C., Warren, B., Dabiré, K. R., Russell, I. J. & Gibson, G. ‘Singing on the wing’ as a mechanism for species recognition in the malarial mosquito Anopheles gambiae. Curr. Biol. 20, 131–136 (2010).CAS 
PubMed 

Google Scholar 
47.Caputo, B. et al. Comparative analysis of epicuticular lipid profiles of sympatric and allopatric field populations of Anopheles gambiae s.s. molecular forms and An. arabiensis from Burkina Faso (West Africa). Insect Biochem. Mol. Biol. 37, 389–398 (2007).CAS 
PubMed 

Google Scholar 
48.Ferguson, H. M. & Read, A. F. Genetic and environmental determinants of malaria parasite virulence in mosquitoes. Proc. R. Soc. B Biol. Sci. 269, 1217–1224 (2002).CAS 

Google Scholar 
49.Niang, A. et al. Semi-field and indoor setups to study malaria mosquito swarming behavior. Parasit. Vectors 12, 446 (2019).PubMed 
PubMed Central 

Google Scholar 
50.Santolamazza, F. et al. Insertion polymorphisms of SINE200 retrotransposons within speciation islands of Anopheles gambiae molecular forms. Malar. J. 7, 163 (2008).
PubMed 
PubMed Central 

Google Scholar 
51.Vantaux, A. et al. Larval nutritional stress affects vector life history traits and human malaria transmission. Sci. Rep. 6, 36778 (2016).CAS 
PubMed 
PubMed Central 
ADS 

Google Scholar 
52.Crawley, M. J. The R Book. (Ltd, Sons, 2012). https://doi.org/10.1002/9780470515075.53.Hothorn, T., Bretz, F., Westfall, P. & Heiberger, R. M. Package ‘multcomp’ title simultaneous inference in general parametric models. Biom. J. 50, 346–363 (2016).
Google Scholar  More

General characterization of seven newly isolated HMO-2011-type phagesIn this study, we used four Roseobacter strains (FZCC0040, FZCC0042, FZCC0012, and FZCC0089) and one SAR11 strain (HTCC1062) to isolate phages. FZCC0040 and FZCC0042 belong to the Roseobacter RCA lineage [22], FZCC0012 shares 99.8% 16S rRNA gene identity with Roseobacter strain HIMB11 [57], and FZCC0089 belongs to a newly identified Roseobacter lineage located close to HIMB11 and SAG-019 lineages (Supplementary Fig. 1).A total of seven phages were newly isolated and analyzed in this study (Table 1). The complete phage genomes range in size from 52.7 to 54.9 kb, harbor 62 to 84 open reading frames (ORFs), and feature a G + C content ranging from 33.8 to 48.6%. Compared to other HMO-2011-type phages, pelagiphage HTVC033P has a relatively lower G + C content of 33.8%, similar to the G + C content of its host HTCC1062 (29.0%) and of other described pelagiphages [21, 26,27,28]. The G + C content of other six roseophages ranges from 42.2 to 48.6%, which is also similar to the G + C content of the hosts they infect (44.8 to 54.1%).Despite their distinct host origins, these phage genomes show considerable similarity in terms of gene content and genome architecture (Fig. 1). They all display clear similarity with the previously reported SAR116 phage HMO-2011 [20] and HMO-2011-type RCA phages [22]. Overall, these phages share 19.2 to 79.1% of their genes with previously reported HMO-2011-type phages and all contain homologues of HMO-2011-type DNA replication and metabolism genes, structural genes, and DNA packaging genes. Moreover, their overall genome structure is conserved with that of HMO-2011-type phages. Considering these observations, we tentatively classified these seven phages into the HMO-2011-type group. Of the 11 currently known HMO-2011-type isolates, one infects the SAR116 strain IMCC1322, one infects the SAR11 strain HTCC1062, and the remaining nine all infect Roseobacter strains; this suggest that HMO-2011-type phages infect diverse bacterial hosts. HTVC033P is the first pelagiphage identified to belong to the HMO-2011-type viral group. Our study has also increased the number of known types of pelagiphages. To date, pelagiphages belonging to a total of nine distinct viral groups have been isolated and analyzed [21, 26,27,28].Fig. 1: Alignment and comparison of genomes of HMO-2011-type isolates and representative HMO-2011-type MVGs from major subgroups.HMO-2011-type phage isolates are shown in red. Phages isolated in this study are indicated with red asterisks. Predicted open reading frames (ORFs) are represented by arrows, with the left or right arrow points indicating the direction of their transcription. The numbers inside the arrows indicate ORF numbers. ORFs annotated with known functions are marked using distinct colors according to their functions. HMO-2011-type core genes are indicated with blue asterisks. The color of the shading connecting homologous genes indicates the level of amino acid identity between the genes. To clearly present the genomic comparison, several MVGs were rearranged to start from the same gene as in the HMO-2011-type phages. DNAP DNA polymerase, Endo endonuclease, RNR ribonucleoside-triphosphate reductase, PhoH phosphate starvation-inducible protein, MazG MazG nucleotide pyrophosphohydrolase domain protein, ThyX thymidylate synthase, GRX glutaredoxin, TerS terminase small subunit, TerL terminase large subunit.Full size imageIdentification and sequence analyses of HMO-2011-type MVGsTo identify HMO-2011-type MVGs, we performed a metagenomic mining and retrieved a total of 207 HMO-2011-type MVGs (≥50% genome completeness) from viromes in the worldwide ocean, from tropical to polar oceans (Supplementary Table 1). These MVGs range in size from 29.2 to 67.9 kb and their G + C content range from 31.3 to 52.4%. In addition, 45 HMO-2011-type MVGs were also identified from some non-marine habitats, suggesting that HMO-2011-type phages are widely distributed worldwide (Supplementary Table 1).Genomic analysis confirmed that all HMO-2011-type MVGs exhibit genomic synteny with HMO-2011-type phages (Fig. 1). Although some of these HMO-2011-type MVGs are highly similar to their cultivated relatives, most MVGs appear to have more genomic variations. To resolve the evolutionary relationship among the HMO-2011-type phages, a phylogenetic tree was constructed based on the concatenated sequences of five core genes. We found that HMO-2011-type phages are evolutionarily diverse and can be separated into at least 10 well-supported subgroups ( >2 members), with 140 MVGs clustering into previously identified HMO-2011-type groups (subgroups I and III in Fig. 2A) [22], and the remaining 67 MVGs forming new subgroups (Fig. 2A). Among these HMO-2011-type subgroups, three contain cultivated representatives (subgroups I, III, and IX). Subgroup I contains the greatest number of phages, including six cultivated representatives and 123 MVGs (Fig. 2A). The cultivated representatives in subgroup I include a phage that infect SAR116 strain and five phages that infect Roseobacter strains. Subgroup III contains four cultivated representatives that infect two Roseobacter strains, and 17 MVGs. Pelagiphage HTVC033P and nine MVGs form subgroup IX. Other subgroups have no cultivated representatives yet. The results of phylogenomic analysis showed that subgroups I to VI are closely related, whereas subgroups VII to X are located on a separate branch and are more distinct from the subgroups I to VI, which suggests that these subgroups are more evolutionarily distant. A phylogenomic-based approach with GL-UVAB workflow [53] was also performed to cluster these HMO-2011-type genomes, which showed similar grouping results (Supplementary Fig. 2).Fig. 2: Phylogenomic and shared-gene analyses of HMO-2011-type phages.A A maximum-likelihood tree was constructed using concatenated sequences of five hallmark genes. HMO-2011-type phages were grouped into 10 subgroups based on the phylogeny. Shading is used to indicate the subgroups. HMO-2011-type phage isolates are shown in red. Genomes containing an integrase gene are indicated by red triangles. The G + C content and completeness of the genomes are indicated. Scale bar indicates the number of amino acid substitutions per site. B Heatmap showing the percentage of shared genes between HMO-2011-type genomes. Phages in the same subgroup are boxed.Full size imageA previous study suggested the use of the percentage of shared proteins as a means of defining phage taxonomic ranks and proposed that phages with ≥20 and ≥40% orthologous proteins in common can be grouped at the taxonomic ranks of subfamily and genus, respectively [58]. Overall, most of the calculated percentages between HMO-2011-type genomes fall within the 20 to 100% range and most of the percentages between genomes within the same subgroup fall within the 40 to 100% range (Fig. 2B). Therefore, our results suggest that the HMO-2011-type is roughly a subfamily-level phage taxonomic group containing at least ten genus-level subgroups in the Podoviridae family.Conserved genomic structure and variation in HMO-2011-type phagesOf the 1235 orthologous protein groups (≥2 members) identified in HMO-2011-type genomes, only 254 proteins groups could be assigned putative biological functions (Supplementary Table 2). Comparative genomic analysis clearly revealed the conserved functional module structure of all HMO-2011-type genomes. All HMO-2011-type phage genomes can be roughly divided into the DNA metabolism and replication module, structural module and DNA packaging module (Fig. 1). Most of the homologous genes are scattered in similar loci of the HMO-2011-type genomes. Core genome analysis based on complete HMO-2011-type genomes revealed that HMO-2011-type genomes share a common set of ten core genes (Fig. 1). These core genes are mostly genes related to essential function in phage replication and development, including genes encoding DNA helicase, DNA primase, DNA polymerase (DNAP), portal protein, capsid protein, and terminase small and large subunits (TerL and TerS) as well as several genes with no known function, suggesting that phages in this group employ similar overall infection and propagation processes (Fig. 1).Most members in subgroups I and III and one member in subgroup II possess a tyrosine integrase gene (int) located upstream of the DNA replication and metabolism module, whereas all subgroup IV to X genomes contain no identifiable lysogeny-related genes. This result suggests that members of subgroups IV to X might be obligate lytic phages. Integrase genes typically occur in the genomes of temperate phages and are responsible for site-specific recombination between phage and host bacterial genomes [59, 60]. In subgroup III, RCA phage CRP-3 has been experimentally demonstrated to be capable of integrating into the host genome [22]. Thus, certain int-containing HMO-2011-type phages are also likely to be temperate phages.In the DNA metabolism and replication modules, genes encoding DNA primase, DNA helicase, DNAP, ribonucleotide reductase (RNR), and endonuclease can be identified; and DNA helicase, DNA primase, and DNAP are core to all HMO-2011-type phages. All reported HMO-2011-type phages contain an atypical DNAP, in which a partial DnaJ central domain is located between the exonuclease domain and the DNA polymerase domain [20, 22]. The Escherichia coli DnaJ protein, a co-chaperone [61], has been shown to be involved in diverse functions [62] and to be critical for the replication of phage Lambda [63,64,65]. The sequence analysis revealed that DNAP sequences of these seven new HMO-2011-type phages and 207 MVGs also present this unusual domain structure and contain two repeats of the CXXCXGXG motifs involved in zinc binding [66] in the partial DnaJ domain (Supplementary Fig. 3). RNR gene is frequently detected in subgroups I, II, III, IV, V, and X genomes but not in the other subgroup genomes. RNRs, which are widely distributed in diverse phage genomes, are involved in catalyzing the reduction of ribonucleotides to deoxyribonucleotides, and thus play a crucial role in providing deoxyribonucleoside triphosphates for phage DNA biosynthesis and repair [67,68,69]. RNR genes clustered with the RNR gene in phage HMO-2011 were previously reported to dominate the class II viral RNRs in examined marine viromes [69]. In the remaining two modules, genes involved in phage structure (e.g., genes encoding capsid and portal proteins), packaging of DNA (TerL and TerS genes), and cell lysis were detected. The proteins encoded by these genes play key roles in phage morphogenesis and virion release.Examination of the distribution of the orthologous groups among the subgroups revealed clear pan-genome differences in various subgroups (Fig. 3). Most subgroups harbor subgroup-specific genes not identified in other subgroups, although  no function has yet been assigned to most of these genes. Notably, the phages in subgroups VII, VIII, and IX possess genomic features that differentiate them from phages in other subgroups, specifically with regard to the G + C content and gene content. The members of these three subgroups are closely related to each other in the phylogenetic tree and harbor several subgroup-specific genes. The G + C content of the phage genomes in these subgroups ranges from 31.9 to 35.4%, significantly smaller than other subgroups but similar to the G + C content of SAR11 bacteria and other known pelagiphages. HTVC033P is the only cultivated representative of subgroup IX. The aforementioned results suggest that the phages in subgroup VII, VIII, and IX might have related bacterial hosts and are highly likely to be pelagiphages. The host prediction using RaFAH tool also assigned Pelagibacter as their potential hosts (Supplementary Table 1). Subgroup X is located near these three subgroups in the phylogenetic tree, and the G + C content of the phages in this subgroup ranges from 34.4 to 39.0%. The host prediction assigned Roseobacter as their potential hosts. The hosts of this subgroup still remain to be experimentally investigated.Fig. 3: Distribution and functional classification of orthologous protein groups across HMO-2011-type genomes.Only orthogroups containing >10 members or showing subgroup-specific features are shown. Subgroup-specific genes are boxed in red. Genes that are absent in a specific subgroup are boxed in orange.Full size imageMetabolic capabilities of HMO-2011-type phagesAll HMO-2011-type phage genomes harbor several host-derived auxiliary metabolic genes (AMGs) potentially involved in diverse metabolic processes. Some AMGs in HMO-2011-type phages have been discussed previously [20, 22].Subgroups VII, VIII, IX, and X possess distinct AMGs as compared with the other subgroups. For example, the genes encoding FAD-dependent thymidylate synthase (ThyX, PF02511) and MazG pyrophosphohydrolase domains are absent in all subgroups VII, VIII, IX, and X genomes but frequently detected in other subgroup genomes. ThyX protein is essential for the conversion of dUMP to dTMP mediated by an FAD coenzyme and is therefore a key enzyme involved in DNA synthesis [70, 71]. The thyX gene is commonly found in microbial genomes and phage genomes. Phage-encoded ThyX has been suggested to compensate for the loss of host-encoded ThyA and thus play crucial roles in phage nucleic acid synthesis and metabolism during infection [72]. Except in the case of subgroups VII, VIII, IX, and X genomes, the mazG gene, which encodes a nucleoside triphosphate pyrophosphohydrolase is sporadically distributed in HMO-2011-type genomes. MazG protein is predicted to be a regulator of nutrient stress and programmed cell death [73] and has been hypothesized to promote phage survival by keeping the host alive during phage propagation [74]. The Escherichia coli MazG can interfere with the function of the MazEF toxin–antitoxin system by decreasing the cellular level of (p)ppGpp [73]. However, a recent study showed that a cyanophage MazG has no binding or hydrolysis activity against alarmone (p)ppGpp but has high hydrolytic activity toward dGTP and dCTP, and it was speculated to play a role in hydrolyzing high G + C host genome for phage replication [75]. Whether the MazG proteins encoded by HMO-2011-type phages play a similar role in phage propagation remained to be investigated.Five MVGs in subgroup I contain a gene encoding a DraG-like family ADP-ribosyl hydrolase (ARH). In cellular ADP-ribosylation systems, ARH catalyzes the cleavage of the ADP-ribose moiety, and thereby counteract the effects of ADP-ribosyl transferases [76]. It has been reported that ARH in Rhodospirillum rubrum regulates the nitrogen fixation [77]. However, the function of this phage-encoded ARH in the phage propagation process remains unclear.We also observed that several MVGs possess genes involved in iron–sulfur (Fe–S) cluster biosynthesis, including an Fe–S cluster assembly scaffold gene (iscU) that involved in Fe–S cluster assembly and transfer [78] and an Fe–S cluster insertion protein gene (erpA). Fe–S cluster participates in a wide variety of cellular biological processes [79]. The discovery of these genes suggests that these phages may play important roles in Fe–S cluster biogenesis and function.The gene encoding sodium-dependent phosphate transport protein (PF02690) has been identified in eight subgroup I genomes. The Na/Pi cotransporter family protein is responsible for high-affinity, sodium-dependent Pi uptake, and thus the protein plays a critical role in maintaining phosphate homeostasis [80]. This gene might function in the transport of phosphate into cells during phage infection. The presence of Na/Pi cotransporter genes suggests that some HMO-2011-type phages may have the potential to regulate host phosphate uptake in phosphate-limited ocean environments in order to benefit phage replication and propagation.Identification and phylogenetic analysis of HMO-2011-type DNAPsThe genetic diversity and geographically distribution of HMO-2011-type phages in marine environments was further inferred from DNAP gene analyses. A total of 2433 HMO-2011-type DNAP sequences with sequence sizes ranging from 540 to 779 amino acids were identified and subjected to phylogenetic analysis (Supplementary Table 3).Among the identified HMO-2011-type DNAPs, 2030 sequences were retrieved from the GOV 2.0 Tara expedition upper-ocean viral populations (0–1000 m), from tropical to polar regions. HMO-2011-type DNAP genes were identified from all analyzed upper-ocean viromes, suggesting the global prevalence of HMO-2011-type phages in upper oceans.A previous study revealed that marine viromes contain various types of tailed phage genomes that encode a family A DNAP gene [81]. To estimate the importance of HMO-2011-type phages, we calculated the proportion of HMO-2011-type DNAPs based on the number of HMO-2011-type DNAP sequences and the total number of family A DNAP sequences ( >470 aa) in each GOV 2.0 viral population dataset. This analysis revealed that HMO-2011-type DNAPs accounted for up to 19.7% of all family A DNAPs in each GOV 2.0 dataset (Supplementary Table 4). We found that the HMO-2011-type DNAP sequences appear to be more dominant in epipelagic viromes than in mesopelagic viromes (p  More