Reconstructing contact and a potential interbreeding geographical zone between Neanderthals and anatomically modern humans

While the interbreeding of Homo neanderthalensis (hereafter Neanderthal) and Anatomically modern human (AMH) has been proven, owing to the shortage of fossils and absence of appropriate DNA, the timing and geography of their interbreeding are not clearly known. In this study, we applied ecological niche modelling (maximum entropy approach) and GIS to reconstruct the palaeodistribution of Neanderthals and AMHs in Southwest Asia and Southeast Europe and identify their contact and potential interbreeding zone during marine isotope stage 5 (MIS 5), when the second wave of interbreeding occurred. We used climatic variables characterizing the environmental conditions of MIS 5 ca. 120 to 80 kyr (averaged value) along with the topography and coordinates of Neanderthal and modern human archaeological sites to characterize the palaeodistribution of each species. Overlapping the models revealed that the Zagros Mountains were a contact and potential interbreeding zone for the two human species. We believe that the Zagros Mountains acted as a corridor connecting the Palearctic/Afrotropical realms, facilitating northwards dispersal of AMHs and southwards dispersal of Neanderthals during MIS 5. Our analyses are comparable with archaeological and genetic evidence collected during recent decades.

There is strong evidence of multiple interbreeding events between two groups of Neanderthals and archaic/ modern humans in western Eurasia e.g. 3,25.7][28] .Palaeogenetic studies have shown that the second wave of interbreeding occurred during MIS 5 4,28 .In some studies, researchers have suggested that the lower latitude regions of southwestern Asia have high potential for the first overlap between Neanderthals and AMHs.Sanchez Goñi 29 examined the patterns of expansion of Neanderthals and AMHs and showed that they shared the same ecological niches under certain climatic conditions during the Late Pleistocene.Recently, Churchill et al. 30 reported facial morphological similarities between Neanderthals and AMHs in the Near East, indicating it could be a key region for interbreeding between the two lineages.However, it is still unclear where the two species met and interbred.
Ecological niche models (ENMs) are very practical tools for investigating the geography of two species' palaeodistribution and potential interbreeding areas 31 .ENMs have been found to have important applications in palaeobiogeography, archaeology and palaeoanthropology [32][33][34][35][36] .They use occurrence data of a target species, including ancient humans, as well as palaeoenvironmental variables to calculate the probability of a species or hominin species' presence in a defined geographic region 37 .These models have successfully been used to reconstruct the distribution of different hominin species 35,36,38 , identify refugia during the ice ages and reconstruct dispersal corridors 39 , niche overlap among species 31 and niche overlap with prey species 40 .For example, Ruan et al. 31 successfully used ENMs to identify the contact zones of Neanderthals and Denisovans.In another study, Benito et al. 36 applied the ecological niches to determine the distribution of Neanderthals during the last interglacial period in Europe and in the Irano-Turonian region.Thus, ENMs can be used to model the palaeodistribution of Neanderthals and AMHs and locate the geography of their niche overlap 31,36 .
The aim of the present study was to reconstruct the palaeodistribution of Neanderthals and AMHs during MIS 5 to identify the contact and potential interbreeding geographical zones of these two species.We also estimated the most important predictor of the two species and investigated the responses of the two species to environmental variables.Previous studies have suggested Southwest Asia as a potential area for the interbreeding of Neanderthals and AMHs 30,41,42 .Notably, this region which is located at the crossroads of the Afrotropical and Palearctic realms 43 , matches the distribution of AMHs and Neanderthals, respectively.Thus, we hypothesized that these two species first met and interbred at the border of these two biogeographic realms where environmental conditions facilitated niche overlap and resource partitioning by providing a highly diverse habitat rich in resources.Climate is a major determinant of species distributions 44 , particularly at large spatial scales, thus we expect climate to be more effective than topography in shaping the interactions between Neanderthals and AMHs.

Reconstructing the contact and interbreeding zone
The models developed in this study for Neanderthals (AUC = 941) and AMHs (AUC = 895) performed well according to the AUC model performance metric.Our model of the palaeodistribution of Neanderthals shows that north and west of the Mediterranean Sea towards the Levant, vast patches in Turkey, around the Black Sea, south of the Caspian Sea, Taurus, Caucasus and Zagros Mountains, were highly suitable for this species during MIS 5 (Fig. 1).The AMH palaeodistribution model identified large and continuous suitable patches in Africa, Arabia and the Iranian Plateau.Our model identified the Zagros Mountains as a contact and potential interbreeding zone in Southwest Asia and Southeast Europe.

Variable importance and response curve
We estimated the relative contributions of the environmental variables to the Maxent model of Neanderthals and AMHs.We found that the maximum temperature of the warmest month (with 58.5% contribution), the minimum temperature of the coldest month (with 19.7% contribution), and the annual precipitation (16.5% contribution) were the most important predictors of the palaeodistribution of Neanderthals 36 .The maximum temperature of the warmest month had a negative association with the presence of Neanderthals.Slope (with 35.6% contribution), topographic diversity (with 26% contribution) and precipitation of the warmest quarter (with 14% contribution) were the most important variables in shaping the palaeodistribution of AMHs.Both species presented similar responses to decreases in slope and habitat suitability in areas with high slopes.Figure 2 shows how each environmental variable affects the Maxent prediction for Neanderthals (a) and AMHs (b).The curves show how the predicted probability of presence changes as each environmental variable is varied, keeping all other environmental variables at their average sample value.

Precipitation changes from 140 to 40 kyr
Figure 3 shows the changes in precipitation from 140 to 40 kyr, with 10,000 intervals for the Zagros Mountains.The highest amount of precipitation occurred at 120 kyr, making it a suitable time for range expansion and interactions between Neanderthals and AMHs.

Discussion
Hominin interbreeding is an important topic in palaeoanthropological studies, but when and where it occurred remain largely unknown.Among the different hominin species, the interbreeding of Neanderthals and AMHs is particularly important, as it contributes to the genetics of our own species.Here, we applied ENMs and GIS and revealed that the Zagros Mountains of Iran is a potentially highly suitable geographic unit for niche overlap and a potential interbreeding zone of these two species.
Our niche models predicted niche overlap for the two species in the Zagros Mountains.In support of this finding, various studies of genetic data e.g. 45,46, ecological modelling e.g. 21, archaeological and genetic records e.g. 24,46and fossils 30,47 are in concordance with our niche overlap model.The expansion of Neanderthals to Zagros must have occurred in accordance with the Palearctic environment and karstic terrains from both sides of the Black Sea, i.e., the Caucasus and Anatolia crossing towards the southern regions.The latest evidence has shown that the southernmost expanse of Neanderthals was up to the latitudes of approximately 31° in an arm-shaped area that stretched to the south in two different directions alongside the Anti-Lebanon and Zagros Mountains 9 .The Neanderthals in territories further east, such as those found in present-day Uzbekistan 47 , Tajikistan 48 and Asian Russia 26,49 , are known as Central and North Asian Neanderthals.To date, evidence of the presence of Neanderthals is consistent with southwestern humid mountainous zones, including Anti-Lebanon in the Levant 50 , Anatolia 51 , the Caucasus 52,53 and Zagros 9,54 .
The data on the MP period in the Zagros Mountains region are rich and more up-to-date due to the discoveries of stratified sites associated with absolute dates, hominin fossil records, and lithic artefacts.Among the large number of MP sites, four yielded Neanderthal fossils.The best-known of these is Shanidar Cave, where the remains of ten Neanderthals were discovered 55 .Approximately 350 km southeast (approximately 34° latitude), the Wezmeh and Bisetun caves in the Kermanshah region also yielded Neanderthal remains 9,54,56 .However, the recent discovery of Neanderthal remains from the Bawa Yawan Rock Shelter is significant since it yielded an     60 and in the southern to central parts of the Persian Plateau 24 .
In accordance with our initial expectation, the interaction and potential interbreeding zone of Neanderthals and AMHs was located in the contact zone of the Afrotropical and Palearctic realms, namely, the Zagros Mountains.There are several reasons why the Zagros Mountains are a suitable location for the niche overlap and potential interbreeding zone of two species.First, the Zagros Mountains are characterized by the environmental conditions of the Palearctic realm, which is the birthplace of Neanderthals 61 .At the same time, the areas surrounding Zagros are characterized by the environmental conditions of the Afrotropical realm, which is the birthplace of AMHs.Thus, the Zagros Mountains could have been visited repeatedly by people living in the border areas of the Palaearctic and Afrotropical realms during Pleistocene climatic shifts.Therefore, the possibility of interactions between different hominins, including AMHs and Neanderthals, was greater in these areas.Second, Zagros covers a vast geographical region (over 1500 km from Lake Van in Turkish Kurdistan to southeastern Iran) capable of supporting large stable human populations.Third, Zagros is exceptionally diverse in terms of topography and biodiversity 40,[62][63][64][65][66] , making it capable of supporting the presence of two species at the same time.These mountains facilitate the niche overlap of some animal species with similar niches within the same habitat 40,67 .These mountains are known to play a very significant role in species distribution by acting as a dispersal barrier or as a dispersal corridor 63,65 .These findings support the results of our study.
Our findings are further supported by new fossil discoveries in the Zagros Mountains and new genetic data 46 .We assume a migration route into the Central Plateau from other directions, including the southern region via Arabia, the Persian Gulf, and the Oman Sea, is plausible 46 .This route might have followed the coastal lines towards the north and eventually into the inner parts of the Persian Plateau.Recent evidence of hominin occupations scattered on the surface in areas located in the southernmost part of the Persian Plateau supports our hypothesis 24,60,68 .
Our initial supposition was that climatic factors would be the predominant force in predicting the distributions of both Neanderthals and AMHs.However, our findings revealed a nuanced picture: while climate emerged as the key determinant of the Neanderthal habitat, AMH distribution was significantly influenced by topographical variations.The climate was homogenous, but the topography was heterogeneous across the AMH distribution areas.These findings likely suggest that topography played a more pronounced role in sculpting the distribution pattern of AMHs.Our study contributes to the growing body of evidence that underscores the complex interplay between environmental factors in determining species distributions.Our results are in line with prey overlap 40 , showing that the annual precipitation and maximum temperature of the warmest month were the most important predictors of Neanderthal distribution on the Persian Plateau.Climate was the most important determinant of Neanderthal distribution in Europe and the Iran-Turanian region during the last interglacial period; however, the influence of topography was confined to local scales 36 .
One particular application of ecological niche models (ENMs) is to identify suitable areas for the presence of target species where no observations have been made 69,70 .Field surveys guided by ENMs have led to the discovery of new populations and rare species 69,70 , thereby proving the utility of ENMs in this context.Our model, which predicts the interbreeding areas of Neanderthals and AMHs, is assigned a very high priority for future field investigations and excavations.Although field testing of ENMs in archaeological studies is limited 40 , we encourage Iranian archaeologists to conduct field excavations in this potential interbreeding area to evaluate the practicality of the models in archaeological research.Moreover, the use of ENMs can guide the allocation of resources for archaeological excavations, ensuring that efforts are concentrated in areas with the highest potential for significant findings.By prioritizing these high-probability locations, researchers can maximize the efficiency of their fieldwork, leading to more targeted and fruitful excavations.

Conclusions
Before this study, our understanding of the interbreeding of AMHs and Neanderthals was based on genetic and morphology data alone 71,72 .For the first time, we applied ENMs as additional and independent lines of information to locate possible geographic locations where the two species interbred.Our study identified the Persian Plateau, particularly the Zagros Mountains, as a potential interbreeding area for AMHs and Neanderthals.The possibility of attracting different hominin groups in the Zagros Mountains is justified by the geographical conditions of this region, since it is located in two different biogeographical zones, namely, the Palearctic and Afrotropical realms.The border areas of two realms are important in biology since they operate as refugia for species from glacial environments.Consequently, some parts of the Zagros Mountains could have been visited repeatedly by people living in the border areas of the Palaearctic and Afrotropical realms during Pleistocene climatic shifts.Therefore, the possibility of interaction between different hominins, including AMHs and Neanderthals, was greater in these areas.
In addition to our findings that the Persian Plateau served as a hub for Homo sapiens after dispersal from Africa 46 , we conclude that this plateau contributed significantly to hominin distribution 40,62,73 , dispersal 24,74,75 and evolution 46,76 , and we await many exciting discoveries that will shed light on human evolution and dispersal.

Archaeological sites
We obtained 38 occurrence points for Neanderthals and 45 for AMHs (Fig. 4), extracted from multiple sources, including the "Role of Culture in Early Expansions of Humans Out of Africa" (ROCEEH: http:// www.roceeh.net) Database (ROAD30,31) and Appendix S1 in Benito et al. 36 .Each archaeological site was associated with one or two species on the basis of fossil records and lithic artefacts.Since our research focuses on the time frame MIS 5 (e.g., 120-80 kyr), we used only the archaeological sites during this period for southwest Asia and www.nature.com/scientificreports/southeast Europe.We carefully examined each coordinate and removed the duplicates.Since the environmental data were at a spatial resolution of ~ 5 km (4.65 km at the equator), we thinned the occurrence data to 5 km to avoid pseudoreplication 40 .This time frame was selected because it is suggested that interbreeding events take place during three different periods 28 .The initial wave of interbreeding occurred ~ 250 to 200 kyr, the second wave of interbreeding occurred ~ 100 to 120 kyr and the third and last interbreeding occurred ~ 60 to 50 kyr.We were unable to find enough archaeological sites associated with the presence of the two species for the first interbreeding event to construct robust niche models; thus, we focused on the second interbreeding event that occurred during MIS 5 28 .

Environmental predictors
We considered environmental variables related to past climate and topography to reconstruct the AMH and Neanderthal niches during MIS 5.As palaeoclimatic variables, we added the maximum temperature of the warmest month, the minimum temperature of the coldest month, the annual precipitation and precipitation of the warmest quarter to the niche models for the MIS 5-time span.Palaeoclimatic data were obtained from Oscillayers, which is a dataset of climatic oscillations over Plio-Pleistocene time scales at high spatial-temporal resolution 77 .We estimated the average values for each of the abovementioned variables during MIS 5 via the raster package v. 3.4-13 78 implemented in the R environment 79 .To consider topography, we included the slope and topographic heterogeneity 36,40 , which were downloaded from EarthEnv 80 .To avoid multicollinearity among the predictors, we calculated a variance inflation factor (VIF; 81 ) via the "vifstep" function in the "usdm" package 82 and ensured that the collinearity among the predictors was low (VIF < 10).

Ecological niche modelling
In this study, we used the maximum entropy modelling approach 83 to reconstruct ecological niche models of Neanderthals and AMHs during MIS 5. Maxent version 3.4.4was used to build the niche models 84 .We used the Maxent model because it has been shown to perform better than other niche modelling methods 37,84 .We then overlapped the two palaeodistribution models to identify potential areas for their contact zones in the QGIS (www.qgis.org).The performance of the niche models was assessed via the area under the curve (AUC) metric of the receiving operator characteristic (ROC) curve 83 .An AUC value of 0.5 indicates that the performance of the model is not better than random, whereas values closer to 1.0 indicate better model performance 85 .The ROC curves were created by selecting 80% of the data for training and 20% for testing.

Figure 1 .
Figure 1.Habitat suitability models of the two Homo species and their potential contact and interbreeding zones in Southwest Asia and Southeast Europe.This figure was generated in QGIS 3.14.1 (www.qgis.org).The figures of the Neanderthal (left) and modern human (right) are adapted from www. demor gen.be.

Figure 2 .
Figure 2. Response curves showing how the presence of Neanderthals (a) and AMHs (b) is related to the environmental variables (Bio5: maximum temperature of the warmest month, Bio6: minimum temperature of the coldest month, Bio12: annual precipitation and Bio18: precipitation of the warmest quarter) (https:// biodi versi tyinf ormat ics.amnh.org/ open_ source/ maxent/).

Figure 4 .
Figure 4. Distribution of key archaeological sites dating between MIS 5 to 3 across southwest Asia and southeast Europe.Map data acquired from http:// www.roceeh.org and created in www.qgis.org.
15 kyr, as shown at the sites of Skhul and Qafzeh58, before the area was permanently occupied by H. sapiens approximately 55,000 years ago59.There is a vast amount of data on hominin (including AMHs) occupations from 400 to 50 kyr in Arabia associated with Eastern African lithic technology see42and references therein, and moreover, physical remains, including AMH finger bones from Al-Wusta dated to ca. 85 kyr15, all indicate that Arabia was a gateway to Eurasia during the middle to late Pleistocene.There is evidence of the presence of nun-Mousterian MP artefacts dating back to 80 kyr in the southern regions of the Persian Plateau, both in the Zagros Vol.:(0123456789) Scientific Reports | (2024) 14:20475 | https://doi.org/10.1038/s41598-024-70206-ywww.nature.com/scientificreports/