Natural selection beyond city limits

Natural selection beyond city limits | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Natural selection beyond city limits Maegwin Bonar, Alexander J. Blumenfeld, Nicole A. Fusco, Leonardo Campagna, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6761695/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Urbanization creates heterogeneous selective landscapes that cause the evolution of urban-rural clines in phenotypic traits. Although cities can introduce novel selection pressures, little attention has been paid to the role of selection outside the city in maintaining urban-rural clines. Here we integrate whole genome sequencing, demographic modeling, and complementary models of selection to test how natural selection in both urban and rural environments shapes the evolution of an urban-rural cline in coat color in eastern gray squirrels ( Sciurus carolinensis ). Coat color polymorphism in this species, which presents as either gray or melanic, is primarily determined by a 24-bp deletion in the melanocortin-1 receptor gene ( Mc1R ). Melanic squirrels are often more prevalent in urban environments but rare or absent in rural forests. Whole genome sequencing and demographic modeling revealed substantially greater urban-rural divergence at Mc1R than the genomic background, suggesting urban-rural clines in melanism are maintained by selection. We applied three separate approaches leveraging demographic and genomic data to estimate the selection coefficient against Mc1R alleles in each habitat, producing a surprising, yet consistent finding: strong selection against the melanic morph in the rural environment and neutrality or near neutrality in the city. Our findings demonstrate that selection outside the city can be sufficient to maintain urban-rural clines, and that urban environments can maintain genetic diversity that would otherwise be lost in rural landscapes. This study provides a rare opportunity to unravel both the spatial dynamics and the selective pressures shaping trait variation in a widespread vertebrate species that thrives across diverse landscapes, highlighting the complex and sometimes protective role of urban landscapes in evolutionary processes. Evolutionary Biology Evolutionary Genetics selection scan demographic modelling selection coefficient urban evolution Sciurus Figures Figure 1 Figure 2 Figure 3 Significance Statement Urban evolution studies often focus on novel selection pressures within cities, but here we show that strong selection outside the city is sufficient to maintain urban-rural clines in coat color in eastern gray squirrels. Our genomic and demographic modeling revealed intense selection against an allele that causes melanism in rural forests, yet near neutrality in the urban core. Our work demonstrates how cities can function as genetic refugia, conserving genetic diversity that is eliminated in rural environments. Introduction The global expansion of urban environments over the past two centuries provides an opportunity to deepen our understanding of rapid evolution in novel habitats. Urbanization dramatically alters environments in many ways, including the increase of impervious cover, pollution, thermal stress, and habitat fragmentation (1). These changes are typically assumed (2) to introduce novel selection pressures that drive the evolution of trait variation (1, 3), including behavior, morphology, and stress tolerance (4–9). Although cities are the primary focus in many urban evolution studies, populations can span across the urbanization gradient occupying both urban and rural habitats that each exert distinct selection pressures. In addition to imposing novel selection pressures, cities may also relax selection pressures historically important in rural habitats (10; Fig. 1 a). The interaction between urban and rural selection pressures, though often overlooked, may play a key role in maintaining trait variation along urbanization gradients (i.e., urban-rural clines). A recent review on how urbanization affects natural selection (10) found that studies estimating selection coefficients to date have focused exclusively on urban habitats, underscoring the need to examine selection across both environments. Here, we used genomic and demographic approaches to test whether selection in urban and rural habitats contributes to the maintenance of coat color clines in eastern gray squirrels ( Sciurus carolinensis ). Squirrels are ubiquitous in deciduous forests of eastern North America and typically occur as one of two color morphs: gray or melanic (black) (11). Historical accounts indicate that the melanic morph – once widespread throughout the northern portion of the species’ range – declined as agricultural intensification and hunting expanded and forest extent diminished in the 1800’s throughout the northeastern United States (12–14); meanwhile gray squirrels of both morphs were introduced in large numbers to city parks starting in the mid-19th century (12). Today, urban-rural clines in coat color are well-documented in the region (15), including our focal city, Syracuse, New York, where the melanic morph is far more prevalent in the city ((16), Figure. 1b). The melanic morph is more conspicuous than the gray morph in secondary forests that dominate the rural landscape today (17), where predation is the primary cause of squirrel mortality (18, 19). In cities, vehicular collisions are the primary cause of squirrel mortality (18). Whilst melanic squirrels may be more susceptible to predation, they are underrepresented by ~ 30% among road-killed squirrels, potentially due to their visual conspicuity or behavioral avoidance (20). A previous study found no difference in survival between color morphs in the city, but greater survival for the gray morph in rural woodlands (16), suggesting the possibility of varying selective regimes in urban and rural habitats. Coat color variation in eastern gray squirrels is shaped by a 24-bp indel in the melanocortin 1 receptor gene ( Mc1R ) (21), with the deletion allele (𝚫24) being incompletely dominant for melanism, such that heterozygotes and homozygotes for the 𝚫24 allele have brown-black to jet black coat color, whereas homozygotes for the wild type allele (+) are gray. The straightforward genetic architecture of this trait makes it an excellent system to track allele frequency change and estimate selection coefficients in the context of broader genomic changes in both urban and rural populations, providing insight into the spatial context of selection across urbanization gradients. Results Population demography and background genomic variation Gray squirrels underwent significant demographic changes over the last 200 years of urbanization, resulting in severe population declines due to habitat loss and over-hunting coincident with human-mediated translocations of squirrels into many cities (12–14, 22). These demographic shifts created potential for strong genetic drift and gene flow to influence genomic variation, making it essential to characterize background population structure to distinguish adaptive from nonadaptive divergence at Mc1R . We sampled 196 squirrels from urban and rural habitats in and around the general Syracuse area (Figure S1) and performed low-coverage whole-genome sequencing (WGS; average 7X) on 46 individuals (Fig. 2 a), and the remaining 150 samples were genotyped for the indel variation at Mc1R using polymerase chain reaction (PCR; Figure S1). Notable genetic structure was identified between urban and rural habitats using the whole genome data (Fig. 2 b). Standard and discriminant principal components analysis showed individuals clustered by urban and rural habitats, with admixed ancestry occurring primarily in the rural environment (Fig. 2 b, Supplemental Results Figure S2). To provide more insight into how drift and gene flow shaped genomic structure between habitats, we fit demographic models with parameters inferred from the joint site frequency spectrum (SFS) via coalescent simulations (23). The strongest model indicated a population bottleneck in the urban habitat with bidirectional migration between habitats (Fig. 2 c, Table S1; see Supplemental Results for details). These findings are consistent with the idea of squirrel introductions in cities (22) and findings from previous studies showing eastern gray squirrels can experience high levels of gene flow across large geographic expanses (24). To gain clearer insights into the degree of genomic differentiation between urban and rural habitats, we characterized genome-wide genetic differentiation using two approaches. First, we estimated F ST along 50kb sliding windows (10kb steps) of each autosome implemented in ANGSD v.0.938 (25) (Fig. 2 e; see Supplemental Results for details). Second, we performed genome-wide scans for significant outlier loci based on principal components analysis (PCA). Using the R (26) package pcadapt (27) we tested how much each SNP was associated with population structure derived from our PCA, assuming that outlier SNPs were indicative of local adaptation to urban and rural environments. All SNPs were regressed against the retained ordination axes (K = 2), and outlier SNPs were selected based on their significant correlation with these axes. This analysis identified 16,158 SNPs with deviating statistics from the null distribution under drift. There were 746 SNPs within 534 windows across the genome shared between the PCA and F ST methods (Figure S4). Many of these significant SNPs were located on or within 50kbp of genes, with 62% of the outliers located on chromosomes 5 (20%) and 6 (42%). We investigated the potential for these regions to be harbors of adaptive function by conducting an analysis for selective sweeps by using the ancestral recombination graph (ARG) to describe genealogical relationships and changes to those relationships along the genome due to historical recombination events (28). Our results suggest that recent selective sweeps tend to contribute more to differentiation between urban and rural habitats at the outlier loci than genetic drift (see Supplemental Results for details). Genetic differentiation at Mc1R relative to the genomic background We used all 196 samples to estimate genetic differentiation between urban and rural populations at the Mc1R indel locus for coat color. Within the urban environment 92 individuals were gray (+/+), 50 were melanic heterozygotes (𝚫24/ +), and 10 were melanic homozygotes (𝚫24/𝚫24). In the rural environment all 44 individuals were gray (+/ +). F ST between urban and rural populations for the Mc1R indel locus was 0.276 (95% CI 0.220–0.333) based on the Weir and Cockerham estimator (29) and a bootstrapping approach. This level of divergence was extreme when compared to the genomic background (Fig. 2 d, one-tailed test; p = 7.835 x 10 −6 ). Our SNP based F ST analysis showed regions surrounding the Mc1R indel had lower divergence than the indel, which underscores the need to analyze the variation at the indel locus directly. We also estimated divergence between color morphs in the urban population to corroborate the Mc1R indel locus as the causal locus underlying the coat color polymorphism. F ST between morphs was 0.588 (95% CI 0.523–0.662) for the Mc1R indel, which was greater than the genomic background (Fig. 2 d, one-tailed test; p < 2.2 x 10 − 16 ). No other locus in the genome had greater divergence between color morphs, yet our sliding window F ST analysis between morphs revealed the second highest window corresponded to a region on chromosome 12 that contains a single gene Smyd3 (Fig. 2 e). In mice, Smyd3 functions in the nucleus, stimulating the transcription of several key regulators involved in cell proliferation, epithelial-mesenchymal transition, and is implicated in cancer development (30, 31). These results support previous findings (21) showing the Mc1R indel is the causal locus for the coat color polymorphism, and presents another region where the connection to coat color expression is not obvious and warrants further investigation. Estimating the selection coefficient for the coat color allele We found much stronger urban-rural divergence at the Mc1R indel locus than the genomic background, pointing to selection playing a significant role in maintaining the urban-rural cline in coat color. To provide insight into the nature of selection in each habitat, we estimated the selection coefficient ( s ) against the Mc1R -𝚫24 allele in urban and rural environments using three independent approaches. Model 1 estimated selection analytically from a model of migration-selection balance (32–34) that assumes the equilibrium frequency difference is determined by the balance between selection removing deleterious alleles and migration introducing deleterious alleles. We used estimates of migration from our demographic model based on WGS data, and we derived contemporary allele frequencies from estimates of the proportion of melanic squirrels in urban and rural environments from standardized field surveys (16). Model 2 compared survival time observations from a translocation experiment (16) to simulations, using Approximate Bayesian Computation (ABC) to estimate the probability of mortality and reproduction, from which we estimated the selection coefficient in urban and rural environments. Model 3 forward simulated allele frequencies using a flexible population model of selection and drift, comprising six parameters that encompassed uncertainty (and prior beliefs) of effective population sizes of both rural and urban populations, generation times, selection strengths, and the ratio of melanics to grays (see Supplemental Methods for full explanation of informative priors). It then used ABC to select the best simulations by comparing observed modern morph frequencies to simulated frequencies. The results of all three models show strong selection against the melanic morph in the rural environment, but no selection in the urban environment. The consistency between all three approaches is noteworthy (Fig. 3 , Table 1 ) and suggests remarkably strong selection in rural forests. Table 1 Summary of selection coefficient estimates from all three independent methods: Model 1: migration-selection balance, Model 2: translocation simulations, Model 3: allele frequency simulations. Source data used to fit model Selection against melanics relative to gray in urban environment Selection against melanics relative to gray in rural environment Model 1 Contemporary allele frequencies in urban and rural populations -0.008 (95% CI = -0.027 to 0.021) 0.086 (95% CI = -0.052 to 0.313) Model 2 Adult survival in urban and rural populations 0.000 (95% CI = -0.128 to 0.127) 0.234 (95% CI = 0.034 to 0.519) Model 3 Forward simulation of allele frequencies in urban and rural populations 0.019 (95% CI = -0.115 to 0.155) 0.109 (95% CI = 0.042 to 0.185) Discussion Our work provides a detailed examination of the adaptive and demographic processes that contribute to the establishment and maintenance of urban-rural clines in trait variation. We found genomic differentiation between urban and rural populations of gray squirrels was significantly greater at Mc1R than the genomic background, suggesting selection plays a role in maintaining coat color clines between urban and rural environments in cities despite significant historical population bottlenecks. By estimating selection coefficients from three independent datasets using three different methodologies, we found strong support for selection against the melanic morph in rural forests, and support for (near) neutrality at the coat color locus in the urban environment. These findings add to our growing understanding of biological evolution across urbanization gradients, emphasizing the role of rural selection in maintaining urban-rural clines and the potential for cities to function as havens that protect and maintain genetic diversity. Detecting very recent selective sweeps with classic population genetic approaches can be challenging. Given that we are examining selection on a very recent time scale, and that the Mc1R -𝚫24 allele predates its introduction to cities (35), the allele may have legacy signatures of molecular evolution prior to urbanization. Forward simulations can be helpful in overcoming this challenge because recent changes to allele frequencies occur on the backdrop of standing variation shaped by historical processes. Our two forward simulation models used different input data: one phenotypic survival estimates, the other modern allele frequencies. These models also made very different prior assumptions about the nature of selection. Despite their differences, each model indicated selection against the melanic morph in the rural environment and neutrality for coat color selection in the urban environment (Fig. 3 ). Estimates of the selection coefficients from an analytical model of migration-selection balance supported the same conclusion. This congruence across three independent approaches – each with different assumptions and data inputs – provides compelling evidence that rural populations experience directional selection for the gray morph, whereas coat color appears to be evolving neutrally or nearly so in the city. The decline in prevalence of melanics in rural environments coincided with a period of extensive forest loss, agricultural intensification, and severe hunting pressure that accompanied the growth of cities. We suspect the altered structure of sparse and therefore lighter regrown forests relative to denser, darker old growth forests made the melanic morph more visible and susceptible to predation, including and especially hunting by people (17). Meanwhile, squirrels were introduced to cities where firearm use - and by extension, hunting - was increasingly prohibited through municipal ordinances for public safety (34, 35), thereby relaxing a source of selection on coat color that remains important in rural forests (18). However, the lack of a signal of selection on coat color in the city does not render novel environmental change in the city unimportant. Our estimated selection coefficients reflect the net outcome of multiple possible constraints on fitness. In the urban environment, the fitness benefits melanic squirrels gain from reduced road mortality may offset costs from other processes (e.g., predation), effectively relaxing the total selection on coat color compared to that in rural environments. A more complete understanding of the evolution of coat color along urbanization gradients will require more investigation of the ecological causes of selection in both urban and rural environments, including the potential for direct and indirect selection via correlated traits (e.g., stress response, energy homeostasis, immunity, and resistance to oxidative stress; (36)). Data from multiple cities are also needed to evaluate the generalities of the findings from this study to clarify whether similar ecological conditions in cities and rural environments alike lead to parallel urban-rural clines, or whether convergence is limited by ecological context and nonadaptive processes (15). In conclusion, our results suggest the high prevalence of squirrel melanism in some cities can be maintained via relaxation of strong selection against the melanic morph outside the city. The mean selection coefficients in the rural population range based on our models ranged from 0.086 to 0.234, reflecting intense selection relative to values typical in other systems (median s = 0.082 across n = 3000 estimates; (37)). Although novel environmental change in cities may play a role in relaxing total selection on coat color, our results challenge the widespread assumption that city environments are the primary source of selection involved in the maintenance of urban-rural clines in trait variation observed across many species. Materials and Methods Sampling and DNA extraction One hundred and ninety-six eastern gray squirrels were collected from urban and rural habitats in and around the general area of Syracuse, New York, USA. Details are provided in Supplementary Methods. Genomic DNA was extracted from a tail tissue or blood as described in (38). Library pools for 46 samples were sent for sequencing at the Yale Center for Genomics, using the NovaSeq 6000 (paired-end 2 x 150 bp reads). DNA from 150 samples were genotyped for the Mc1R -𝚫24 allele using polymerase chain reaction (PCR). Whole genome raw data processing and SNP genotyping Raw sequences were cleaned and aligned to the S. carolinensis reference genome (39) (see details in Supplementary Methods). For low and medium coverage sequence data (< 20X), using genotype likelihoods is generally recommended over called genotypes (40) given that lower coverage may increase the influence of sequencing or mapping errors. Therefore, we estimated genotype likelihoods using the SAMtools modified MAQ model (41) implemented in ANGSD v.0.938 (25) and used for most of the following analyses (unless otherwise specified). However, certain analyses still require the use of called genotypes; in these cases, we used ANGSD to directly call genotypes. To ensure the robustness of results generated from called genotypes, we reran all analyses performed using genotype likelihoods (aside from the initial linkage disequilibrium estimates) with called genotypes and checked for significant deviations. Additionally, we excluded the sex chromosomes from each of the following analyses to minimize the potential impact of sex differences on results. The full specification of ANGSD parameters used for SNP genotyping, as well as the code for each of the following analyses, can be found at https://github.com/agentzero93/syr_squirrel_lcwgs . After sequencing and read trimming, the 46 lcWGS samples ranged from 80 to 155 million paired-end reads (mean = 121 million). Alignment to the S. carolinensis reference genome (39) was highly successful for all samples (range 91.3% − 94.7%; mean = 92.7%), thus all samples were retained for variant calling. A total of 2,180,347,754 sites were identified across 19 autosomes for the 46 samples. Prior to examining population structure between morphs (i.e., melanic vs. gray coat color) and habitats (i.e., urban vs. rural), we filtered SNPs for linkage disequilibrium using ngsLD (42), details in Supplemental Methods. Following pruning based on linkage disequilibrium, the pruned dataset which was utilized for all downstream analyses contained a total of 2,297,744 SNPs (6.5X mean coverage and 1.21% mean missing data). MC1R-𝚫24 allele genotyping To further explore the genetic differentiation associated with the Mc1R -𝚫24 allele, we determined the genotype of all samples. The 46 samples with whole-genome sequencing were assessed visually for the Mc1R -𝚫24 allele in Integrative Genomics Viewer (43) by comparing the cleaned raw sequences to the reference genome. The remaining 150 samples were genotyped for the Mc1R -𝚫24 allele using polymerase chain reaction (PCR), details in Supplementary Methods. Population demography To understand population structure and demography of eastern gray squirrels within urban and rural habitats, we first assess if the sample set included closely related individuals (for more information see Supplementary Methods). After confirming individuals were unrelated, we performed an admixture and covariance matrix estimation using PCAngsd (44). The estimated covariance matrix was used as the basis for a principal component analysis (PCA), using the dudi.pca function from the ade4 R package (45). Both the PCA and admixture plots (46) were constructed using R. We tested for isolation-by-distance (IBD) with a Mantel test (47) between a matrix of genetic distances and a matrix of geographic distances using adegenet (48) (see Supplemental Results Figure S8). We calculated 𝝅 (nucleotide diversity), and Tajima’s D (49) using ANGSD in 5kb windows with 1kb steps across all autosomes (see Supplemental Results Figure S9). We explicitly evaluated several demographic scenarios for urban and rural habitats to explore the demographic history for each habitat. We performed population modeling by inferring demographic parameters from the joint site frequency spectrum (SFS) using coalescent simulations using fastsimcoal2 (23), details in Supplementary Methods. Genetic differentiation We estimated population differentiation between habitats (urban vs. rural) and morphs (melanic vs. gray) across the whole genome using two approaches: 1) we estimated genome wide F ST and; 2) genetic differentiation using PCA; and identified outlier loci for subsequent analyses, (details in Supplementary Methods). For the Mc1R -𝚫24 allele we estimated F ST using the Weir and Cockerham estimator (29) and a bootstrapping approach, (details in Supplementary Methods). Signatures of selective sweeps To identify selective sweeps within outlier regions of interest in both habitat and morph comparisons, we followed the protocol outlined in (28) which uses the ancestral recombination graph (ARG) that describes genealogical relationships and changes to those relationships along the genome due to historical recombination events, (details in Supplementary Methods). Model 1: migration-selection balance For the model of migration-selection balance (32–34), the change in the deleterious allele frequency in a population is given by: $$\:\varDelta\:q=\frac{-spq[q+h(p-q\left)\right]}{1-sq(2hp+q})+mQ-Mq$$ In this model, there are two components driving allele frequency change. The first is selection (s), and its efficacy is influenced by the relative allele frequencies (p and q) and the dominance coefficient (h). The second is migration, which introduces deleterious alleles based on the immigration rate and the deleterious allele frequency outside the population (mQ). Emigration of deleterious alleles can also occur and is based on the migration rate out of the population and the deleterious allele frequency in the population (Mq). We used observed values of q and Q derived from estimates of the proportion of melanic squirrels in the urban and rural environments in (16). We drew from the posterior distribution of for the proportion melanic from the clinal model and used estimates of m, M, from the demographic model generated using our 46 whole genome samples. We report the results using the dominance coefficient h = 1 however, we estimated selection for h = 0.7,0.8, and 0.9 as well given that the Mc1R -𝚫24 allele shows incomplete dominance (Figure S10). Using a bootstrapping approach, (details in Supplementary Methods) we estimated by setting Δq equal to 0 and solved for s. Model 2: Translocation simulations Our translocation data provided the day of death for 76 adult squirrels, comprising melanic and gray, released into either an urban or rural environment. We generated 10 million simulations of 76 agents subjected to a daily death probability sampled from a prior uniform distribution of 0 to 0.4, and used an Approximate Bayesian Computation (ABC) approach to select the best 1000 simulations that most closely matched the death days of the observed data.This provided posterior estimates of the daily death probability for rural melanics (0.0159, 95% CI = 0.0102 to 0.0225), urban melanics (0.0060, 95% CI = 0.0038 to 0.0086), rural grays (0.0061, 95% CI = 0.0038 to 0.0088), and urban grays (0.0060, 95% CI = 0.0039 to 0.0084). Kolmogorov-Smirnov tests revealed no difference between urban melanics, rural grays and urban grays, but a significant difference between these three and rural melanics (p < 2.2x10 − 16 ). We then converted the daily death probabilities into reproductive probabilities by assuming reproduction occured at some point between day 5 and day 50 for those that survived. Finally, we converted the reproductive probabilities into selection coefficient estimates ( s ) as: where r is the relative reproductive probabilities (see Supplemental Methods Model 2). These were then converted to positive values for simulations where r 2 belonged to grays, and negative values where r 2 belonged to melanics, thus ensuring the interpretation of the selection coefficient is always with respect to grays. Model 3: Allele frequency simulations. We forward simulated allele frequencies using a flexible population model of selection and drift (see Figure S5). It modelled a rural population of gray and melanic squirrels starting at 1700 CE (prior to urbanization, with an effective population size sampled from 100 to 100,000 (parameter Nr) using a log uniform distribution to favor smaller population sizes (ensuring uncertainty from the effects of drift were well represented). Modern observed allele frequencies were obtained from published literature (16) (frequency of grays in rural = 0.9554, frequency of grays in urban = 0.5248), and after generating 1 billion simulations under this model, we used ABC to skim the best 5000 simulations, using the sum of the squared differences between simulated and observed modern allele frequencies. Declarations Author Contributions: The authors confirm contribution to the paper as follows: study conception and design : M Bonar, AJ Blumenfeld, A Caccone, BJ Cosentino, JP Gibbs, L Campagna; data collection : NA Fusco, M Bonar, AJ Blumenfeld, BJ Cosentino, JP Gibbs; analysis and interpretation of results : M Bonar, AJ Blumenfeld, A Timpson, M G Thomas, A Caccone; writing of the draft manuscript : M Bonar; major comments and edits provided for the draft manuscript : NA Fusco, L Campagna, A Caccone, BJ Cosentino, JP Gibbs, AJ Blumenfeld, A Timpson, M G Thomas. All authors reviewed the results and approved the final version of the manuscript. Competing Interest Statement: The authors declare no competing interest. Acknowledgments Support for this research was provided by the National Science Foundation (DEB 2017987, DEB 2018140, DEB 2018249). We thank J. Proctor, R. Rich, J. Tooley, and J. Vanek for assistance collecting samples. References N. B. Grimm, et al. , Global change and the ecology of cities. Science 319 , 756–760 (2008). M. Alberti, et al. , Global urban signatures of phenotypic change in animal and plant populations. Proc. Natl. Acad. Sci. 114 , 8951–8956 (2017). M. T. J. Johnson, J. Munshi-South, Evolution of life in urban environments. Science 358 , eaam8327 (2017). S. W. Breck, S. A. Poessel, P. Mahoney, J. K. Young, The intrepid urban coyote: a comparison of bold and exploratory behavior in coyotes from urban and rural environments. Sci. Rep. 9 , 2104 (2019). T. Merckx, et al. , Urbanization extends flight phenology and leads to local adaptation of seasonal plasticity in Lepidoptera . Proc. Natl. Acad. Sci. 118 , e2106006118 (2021). H. Eggenberger, et al. , Urban bumblebees are smaller and more phenotypically diverse than their rural counterparts. J. Anim. Ecol. 88 , 1522–1533 (2019). K. J. Parsons, et al. , Skull morphology diverges between urban and rural populations of red foxes mirroring patterns of domestication and macroevolution. Proc. R. Soc. B Biol. Sci. 287 , 20200763 (2020). K. I. Brans, R. A. Almeida, M. Fajgenblat, Genetic differentiation in pesticide resistance between urban and rural populations of a nontarget freshwater keystone interactor, Daphnia magna . Evol. Appl. 14 , 2541–2552 (2021). S. C. Campbell-Staton, et al. , Parallel selection on thermal physiology facilitates repeated adaptation of city lizards to urban heat islands. Nat. Ecol. Evol. 4 , 652–658 (2020). A. Charmantier, et al. , How does urbanization affect natural selection? Funct. Ecol. 1365-2435.14667 (2024). https://doi.org/10.1111/1365-2435.14667. J. L. Koprowski, K. E. Munroe, A. J. Edelman, “Gray not grey: ecology of Sciurus carolinensis in their native range in North America” in Grey Squirrels: Ecology and Management of an Invasive Species in Europe , C. M. Shuttleworth, et al. , Ed. (European Squirrel Initiative, 2016). D. L. Allen, Michigan fox squirrel management (Michigan Dept. Of Conservation, 1943). A. W. Schorger, “Squirrels in early Wisconsin” in Transactions of the Wisconsin Academy of Sciences, Arts, and Letters , (1949). G. I. Robertson, “Distribution of color morphs of Sciurus carolinensis in eastern North America,” University of Western Ontario. (1973). B. J. Cosentino, J. P. Gibbs, Parallel evolution of urban–rural clines in melanism in a widespread mammal. Sci. Rep. 12 , 1752 (2022). B. J. Cosentino, J. P. Vanek, J. P. Gibbs, Rural selection drives the evolution of an urban–rural cline in coat color in gray squirrels. Ecol. Evol. 13 , e10544 (2023). J. Proctor, A. Bryan, B. J. Cosentino, J. P. Gibbs, Crypsis in a polymorphic mammal along an urbanization gradient. Urban Ecosyst. 28 , 94 (2025). R. A. Mccleery, Changes in fox squirrel anti-predator behaviors across the urban–rural gradient. Landsc. Ecol. 24 , 483–493 (2009). M. A. Bowers, B. Breland, Foraging of gray squirrels on an urban‐rural gradient: use of the GUD to assess anthropogenic impact. Ecol. Appl. 6 , 1135–1142 (1996). A. F. Parlin, et al. , Road mortality contributes to the evolution of an urban–rural cline in squirrel coat color. Evol. Appl. 18 , e70109 (2025). H. McRobie, A. Thomas, J. Kelly, The genetic basis of melanism in the gray squirrel ( Sciurus carolinensis ). J. Hered. 100 , 709–714 (2009). E. Benson, The urbanization of the eastern gray squirrel in the United States. J. Am. Hist. 100 , 691–710 (2013). L. Excoffier, et al. , fastsimcoal2 : demographic inference under complex evolutionary scenarios. Bioinformatics 37 , 4882–4885 (2021). N. A. Fusco, et al. , Population genomic structure of a widespread, urban‐dwelling mammal: The eastern grey squirrel ( Sciurus carolinensis ). Mol. Ecol. 33 , e17230 (2024). T. S. Korneliussen, A. Albrechtsen, R. Nielsen, ANGSD: analysis of next generation sequencing data. BMC Bioinformatics 15 , 356 (2014). R Core Team, R: A language and environment for statistical computing. (2022). Deposited 2022. K. Luu, E. Bazin, M. G. B. Blum, pcadapt : an R package to perform genome scans for selection based on principal component analysis. Mol. Ecol. Resour. 17 , 67–77 (2017). H. A. Hejase, et al. , Genomic islands of differentiation in a rapid avian radiation have been driven by recent selective sweeps. Proc. Natl. Acad. Sci. 117 , 30554–30565 (2020). B. Weir, C. Clark Cockerham, Estimating F-statistics for the analysis of population structure. Evolution 38 , 1358–1370 (1984). M. E. Sarris, P. Moulos, A. Haroniti, A. Giakountis, I. Talianidis, Smyd3 is a transcriptional potentiator of multiple cancer-promoting genes and required for liver and colon cancer development. Cancer Cell 29 , 354–366 (2016). D. Yang, et al. , H3K4 methyltransferase Smyd3 mediates vascular smooth muscle cell proliferation, migration, and neointima formation. Arterioscler. Thromb. Vasc. Biol. 41 , 1901–1914 (2021). J. B. S. Haldane, A mathematical theory of natural and artificial selection. (Part VI, Isolation.). Math. Proc. Camb. Philos. Soc. 26 , 220–230 (1930). S. Wright, Evolution in Mendelian populations. Genetics 16 , 97–159 (1931). H. E. Hoekstra, K. E. Drumm, M. W. Nachman, Ecological genetics of adaptive color polymorphism in pocket mice: geographic variation in selected and neutral genes. Evolution 58 , 1329–1341 (2004). H. R. McRobie, N. D. Moncrief, N. I. Mundy, Multiple origins of melanism in two species of North American tree squirrel ( Sciurus ). BMC Evol. Biol. 19 , 140 (2019). A. Roulin, A.-L. Ducrest, Association between melanism, physiology and behaviour: a role for the melanocortin system. Eur. J. Pharmacol. 660 , 226–233 (2011). T. J. Thurman, R. D. H. Barrett, The genetic consequences of selection in natural populations. Mol. Ecol. 25 , 1429–1448 (2016). N. A. Fusco, E. Pehek, J. Munshi‐South, Urbanization reduces gene flow but not genetic diversity of stream salamander populations in the New York City metropolitan area. Evol. Appl. 14 , 99–116 (2021). D. Mead, et al. , The genome sequence of the eastern grey squirrel, Sciurus carolinensis Gmelin, 1788. Wellcome Open Res. 5 , 27 (2020). R. Nielsen, J. S. Paul, A. Albrechtsen, Y. S. Song, Genotype and SNP calling from next-generation sequencing data. Nat. Rev. Genet. 12 , 443–451 (2011). H. Li, et al. , The sequence alignment/map format and SAMtools. Bioinformatics 25 , 2078–2079 (2009). E. A. Fox, A. E. Wright, M. Fumagalli, F. G. Vieira, ngsLD : evaluating linkage disequilibrium using genotype likelihoods. Bioinformatics 35 , 3855–3856 (2019). J. T. Robinson, et al. , Integrative genomics viewer. Nat. Biotechnol. 29 , 24–26 (2011). J. Meisner, A. Albrechtsen, Inferring Population Structure and Admixture Proportions in Low-Depth NGS Data. Genetics 210 , 719–731 (2018). S. Dray, A.-B. Dufour, The ade4 package: implementing the duality diagram for ecologists. J. Stat. Softw. 22 (2007). R. M. Francis, pophelper : an R package and web app to analyse and visualize population structure. Mol. Ecol. Resour. 17 , 27–32 (2017). P. E. Smouse, J. C. Long, R. R. Sokal, Multiple regression and correlation extensions of the mantel test of matrix correspondence. Syst. Zool. 35 , 627 (1986). T. Jombart, adegenet : a R package for the multivariate analysis of genetic markers. Bioinformatics 24 , 1403–1405 (2008). F. Tajima, Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 595 , 585–595 (1989). Supplementary Tables Supplementary Tables are not available with this version Additional Declarations The authors declare no competing interests. Supplementary Files Bonarsquirrelwgssuppmat.docx Supplementary Information Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6761695","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":462670286,"identity":"c5a444d1-441b-4341-865b-e76e61210ad3","order_by":0,"name":"Maegwin Bonar","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYNACAzACggoGBj4QzUO8ljMMDGzEaWGAamFsI0KLbvvhw695Cu4wmLP3HvxcOe+wPBt7A+ODt224tZidSUuz5jF4xmDZcy5Z8uy2w4ZtPAeYDefi03Igx8w4x+Awg8GNHAPJxm2HGdskEtikefFpOf/+G0yL8c/GOYft2+QfsP/Gq+VGDvNjqBYzycaGw4ltEgxszPi1PDNj/mNwmMey54yZZcOx9OQ2nsRmyTnn8Dks+fHHGX8Oy5mz9xjfbKixtu1nP3zww5sy3FqAgE2CARERzUDM2IBXPRAwf0Di1BFSPQpGwSgYBSMQAAB4kVIr93tg0wAAAABJRU5ErkJggg==","orcid":"","institution":"Yale University","correspondingAuthor":true,"prefix":"","firstName":"Maegwin","middleName":"","lastName":"Bonar","suffix":""},{"id":462670695,"identity":"7cdf33ad-ea98-4889-847d-8ba63550d488","order_by":1,"name":"Alexander J. Blumenfeld","email":"","orcid":"","institution":"Yale University","correspondingAuthor":false,"prefix":"","firstName":"Alexander","middleName":"J.","lastName":"Blumenfeld","suffix":""},{"id":462673932,"identity":"5b8f6823-0cd9-4558-ae81-98405f510b26","order_by":2,"name":"Nicole A. Fusco","email":"","orcid":"","institution":"University of Connecticut Stamford","correspondingAuthor":false,"prefix":"","firstName":"Nicole","middleName":"A.","lastName":"Fusco","suffix":""},{"id":462673933,"identity":"edb53b6c-7559-4c5e-8fe1-1aa633bb653e","order_by":3,"name":"Leonardo Campagna","email":"","orcid":"","institution":"Cornell University","correspondingAuthor":false,"prefix":"","firstName":"Leonardo","middleName":"","lastName":"Campagna","suffix":""},{"id":462673934,"identity":"1450aa89-478f-47c7-a8c1-91c711d14d75","order_by":4,"name":"Adrian Timpson","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Adrian","middleName":"","lastName":"Timpson","suffix":""},{"id":462673935,"identity":"9bdea2b5-0b2d-4d3a-8ae0-6fa02b7ca595","order_by":5,"name":"Mark G. Thomas","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Mark","middleName":"G.","lastName":"Thomas","suffix":""},{"id":462673936,"identity":"f68f824f-0b92-4b69-aa8a-2c3eef0a287d","order_by":6,"name":"Bradley J. Cosentino","email":"","orcid":"","institution":"Hobart and William Smith Colleges","correspondingAuthor":false,"prefix":"","firstName":"Bradley","middleName":"J.","lastName":"Cosentino","suffix":""},{"id":462673937,"identity":"8f8f541e-9741-4bf8-bd48-18f7c2eb89d1","order_by":7,"name":"James P. Gibbs","email":"","orcid":"","institution":"State University of New York College of Environmental Science and Forestry","correspondingAuthor":false,"prefix":"","firstName":"James","middleName":"P.","lastName":"Gibbs","suffix":""},{"id":462673938,"identity":"d6792748-a309-4b50-bdf6-f469be05ba10","order_by":8,"name":"Adalgisa Caccone","email":"","orcid":"","institution":"Yale University","correspondingAuthor":false,"prefix":"","firstName":"Adalgisa","middleName":"","lastName":"Caccone","suffix":""}],"badges":[],"createdAt":"2025-05-27 17:36:07","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":true,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":true},"doi":"10.21203/rs.3.rs-6761695/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6761695/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":83628589,"identity":"1e7c708b-fce5-4219-b1cb-cc7f364e85e9","added_by":"auto","created_at":"2025-05-29 17:54:20","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":776298,"visible":true,"origin":"","legend":"\u003cp\u003e(A) Hypotheses for how urbanization influences natural selection on phenotypic traits resulting in urban-rural clines. (B) Urban-rural cline for coat color of eastern gray squirrels (\u003cem\u003eSciurus carolinensis\u003c/em\u003e) in Syracuse, New York, USA. The left panel shows pie charts indicating the proportion of melanic (black) individuals among 76 sites inferred from point count surveys and camera traps. The right panel shows the estimated proportion of melanic individuals and the 95% credible interval in relation to the distance from the city center. Data redrawn from \u003ca href=\"https://www.zotero.org/google-docs/?jup9kt\"\u003e(16)\u003c/a\u003e.\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-6761695/v1/f46d39e40c87aa22cf2754f8.png"},{"id":83628588,"identity":"b3f96391-0d28-455d-8c23-41955e2ae24f","added_by":"auto","created_at":"2025-05-29 17:54:20","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3246645,"visible":true,"origin":"","legend":"\u003cp\u003e(A) Sites where 46 gray squirrels were sampled from the urban (purple) and rural (blue) areas associated with Syracuse, New York, USA. Circles represent the urban melanic morph (n = 14), squares represent the urban gray morph (n = 12), and triangles represent the rural gray morph (n = 20). (B) Top panel shows a principal component analysis (PCA) using all 46 individuals, and the bottom panel shows admixture proportions (best K = 2). (C) The demographic model best fitting the data which suggests a population bottleneck in the urban habitat with two-way migration between urban and rural habitats. (D) Top panel shows \u003cem\u003eF\u003c/em\u003e\u003csub\u003e\u003cem\u003eST\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e \u003c/em\u003ecalculated using 196 genotyped individuals for the 24bp-deletion at \u003cem\u003eMc1R\u003c/em\u003e for the urban vs. rural habitat comparison. Bottom panel shows the \u003cem\u003eF\u003c/em\u003e\u003csub\u003e\u003cem\u003eST \u003c/em\u003e\u003c/sub\u003ecalculated for the melanic vs. gray comparison only using individuals from the urban habitat. (E) Top panel shows whole genome genetic differentiation (\u003cem\u003eF\u003c/em\u003e\u003csub\u003e\u003cem\u003eST\u003c/em\u003e\u003c/sub\u003e 50kb windows) between urban and rural habitats. Bottom panels show genetic differentiation between melanic and gray morphs within the urban habitat.\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-6761695/v1/58b1b7c0ade89d21f2a02da2.png"},{"id":83628590,"identity":"0b7d62c1-f86f-4fa2-967a-bca64c1d097b","added_by":"auto","created_at":"2025-05-29 17:54:20","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":952181,"visible":true,"origin":"","legend":"\u003cp\u003eEstimates of the strength of selection against the\u003cem\u003e Mc1R\u003c/em\u003e-𝚫24 allele relative to the allele without the deletion in urban (A) and rural (B) habitats for three independent approaches: Model 1: migration-selection balance, Model 2: translocation simulations, Model 3: allele frequency simulations.\u003c/p\u003e","description":"","filename":"Fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-6761695/v1/b5ec15fc145d576bd307fccf.png"},{"id":83629138,"identity":"5da827fa-95b3-454c-9d75-e1b26238e19e","added_by":"auto","created_at":"2025-05-29 18:10:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5190927,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6761695/v1/c8e9cebc-c4e1-448a-be20-4a61b36fb52c.pdf"},{"id":83628592,"identity":"fdc060f6-0645-4580-aa5d-647ffcb843d6","added_by":"auto","created_at":"2025-05-29 17:54:20","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":4093976,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Information\u003c/p\u003e","description":"","filename":"Bonarsquirrelwgssuppmat.docx","url":"https://assets-eu.researchsquare.com/files/rs-6761695/v1/b6c7d73dc8c4aa88505f316f.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eNatural selection beyond city limits\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Significance Statement","content":"\u003cp\u003eUrban evolution studies often focus on novel selection pressures within cities, but here we show that strong selection outside the city is sufficient to maintain urban-rural clines in coat color in eastern gray squirrels. Our genomic and demographic modeling revealed intense selection against an allele that causes melanism in rural forests, yet near neutrality in the urban core. Our work demonstrates how cities can function as genetic refugia, conserving genetic diversity that is eliminated in rural environments.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003eThe global expansion of urban environments over the past two centuries provides an opportunity to deepen our understanding of rapid evolution in novel habitats. Urbanization dramatically alters environments in many ways, including the increase of impervious cover, pollution, thermal stress, and habitat fragmentation (1). These changes are typically assumed (2) to introduce novel selection pressures that drive the evolution of trait variation (1, 3), including behavior, morphology, and stress tolerance (4\u0026ndash;9). Although cities are the primary focus in many urban evolution studies, populations can span across the urbanization gradient occupying both urban and rural habitats that each exert distinct selection pressures. In addition to imposing novel selection pressures, cities may also relax selection pressures historically important in rural habitats (10; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). The interaction between urban and rural selection pressures, though often overlooked, may play a key role in maintaining trait variation along urbanization gradients (i.e., urban-rural clines). A recent review on how urbanization affects natural selection (10) found that studies estimating selection coefficients to date have focused exclusively on urban habitats, underscoring the need to examine selection across both environments.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eHere, we used genomic and demographic approaches to test whether selection in urban and rural habitats contributes to the maintenance of coat color clines in eastern gray squirrels (\u003cem\u003eSciurus carolinensis\u003c/em\u003e). Squirrels are ubiquitous in deciduous forests of eastern North America and typically occur as one of two color morphs: gray or melanic (black) (11). Historical accounts indicate that the melanic morph \u0026ndash; once widespread throughout the northern portion of the species\u0026rsquo; range \u0026ndash; declined as agricultural intensification and hunting expanded and forest extent diminished in the 1800\u0026rsquo;s throughout the northeastern United States (12\u0026ndash;14); meanwhile gray squirrels of both morphs were introduced in large numbers to city parks starting in the mid-19th century (12). Today, urban-rural clines in coat color are well-documented in the region (15), including our focal city, Syracuse, New York, where the melanic morph is far more prevalent in the city ((16), Figure. 1b). The melanic morph is more conspicuous than the gray morph in secondary forests that dominate the rural landscape today (17), where predation is the primary cause of squirrel mortality (18, 19). In cities, vehicular collisions are the primary cause of squirrel mortality (18). Whilst melanic squirrels may be more susceptible to predation, they are underrepresented by ~\u0026thinsp;30% among road-killed squirrels, potentially due to their visual conspicuity or behavioral avoidance (20). A previous study found no difference in survival between color morphs in the city, but greater survival for the gray morph in rural woodlands (16), suggesting the possibility of varying selective regimes in urban and rural habitats.\u003c/p\u003e \u003cp\u003eCoat color variation in eastern gray squirrels is shaped by a 24-bp indel in the melanocortin 1 receptor gene (\u003cem\u003eMc1R\u003c/em\u003e) (21), with the deletion allele (\u0026#120491;24) being incompletely dominant for melanism, such that heterozygotes and homozygotes for the \u0026#120491;24 allele have brown-black to jet black coat color, whereas homozygotes for the wild type allele (+) are gray. The straightforward genetic architecture of this trait makes it an excellent system to track allele frequency change and estimate selection coefficients in the context of broader genomic changes in both urban and rural populations, providing insight into the spatial context of selection across urbanization gradients.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePopulation demography and background genomic variation\u003c/h2\u003e \u003cp\u003eGray squirrels underwent significant demographic changes over the last 200 years of urbanization, resulting in severe population declines due to habitat loss and over-hunting coincident with human-mediated translocations of squirrels into many cities (12\u0026ndash;14, 22). These demographic shifts created potential for strong genetic drift and gene flow to influence genomic variation, making it essential to characterize background population structure to distinguish adaptive from nonadaptive divergence at \u003cem\u003eMc1R\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eWe sampled 196 squirrels from urban and rural habitats in and around the general Syracuse area (Figure S1) and performed low-coverage whole-genome sequencing (WGS; average 7X) on 46 individuals (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea), and the remaining 150 samples were genotyped for the indel variation at \u003cem\u003eMc1R\u003c/em\u003e using polymerase chain reaction (PCR; Figure S1). Notable genetic structure was identified between urban and rural habitats using the whole genome data (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). Standard and discriminant principal components analysis showed individuals clustered by urban and rural habitats, with admixed ancestry occurring primarily in the rural environment (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb, Supplemental Results Figure S2). To provide more insight into how drift and gene flow shaped genomic structure between habitats, we fit demographic models with parameters inferred from the joint site frequency spectrum (SFS) via coalescent simulations (23). The strongest model indicated a population bottleneck in the urban habitat with bidirectional migration between habitats (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec, Table S1; see Supplemental Results for details). These findings are consistent with the idea of squirrel introductions in cities (22) and findings from previous studies showing eastern gray squirrels can experience high levels of gene flow across large geographic expanses (24).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo gain clearer insights into the degree of genomic differentiation between urban and rural habitats, we characterized genome-wide genetic differentiation using two approaches. First, we estimated \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e along 50kb sliding windows (10kb steps) of each autosome implemented in ANGSD v.0.938 (25) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee; see Supplemental Results for details). Second, we performed genome-wide scans for significant outlier loci based on principal components analysis (PCA). Using the R (26) package pcadapt (27) we tested how much each SNP was associated with population structure derived from our PCA, assuming that outlier SNPs were indicative of local adaptation to urban and rural environments. All SNPs were regressed against the retained ordination axes (K\u0026thinsp;=\u0026thinsp;2), and outlier SNPs were selected based on their significant correlation with these axes. This analysis identified 16,158 SNPs with deviating statistics from the null distribution under drift. There were 746 SNPs within 534 windows across the genome shared between the PCA and \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e methods (Figure S4). Many of these significant SNPs were located on or within 50kbp of genes, with 62% of the outliers located on chromosomes 5 (20%) and 6 (42%). We investigated the potential for these regions to be harbors of adaptive function by conducting an analysis for selective sweeps by using the ancestral recombination graph (ARG) to describe genealogical relationships and changes to those relationships along the genome due to historical recombination events (28). Our results suggest that recent selective sweeps tend to contribute more to differentiation between urban and rural habitats at the outlier loci than genetic drift (see Supplemental Results for details).\u003c/p\u003e \u003cp\u003e \u003cb\u003eGenetic differentiation at\u003c/b\u003e \u003cb\u003eMc1R\u003c/b\u003e \u003cb\u003erelative to the genomic background\u003c/b\u003e\u003c/p\u003e \u003cp\u003eWe used all 196 samples to estimate genetic differentiation between urban and rural populations at the \u003cem\u003eMc1R\u003c/em\u003e indel locus for coat color. Within the urban environment 92 individuals were gray (+/+), 50 were melanic heterozygotes (\u0026#120491;24/ +), and 10 were melanic homozygotes (\u0026#120491;24/\u0026#120491;24). In the rural environment all 44 individuals were gray (+/ +). \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e between urban and rural populations for the \u003cem\u003eMc1R\u003c/em\u003e indel locus was 0.276 (95% CI 0.220\u0026ndash;0.333) based on the Weir and Cockerham estimator (29) and a bootstrapping approach. This level of divergence was extreme when compared to the genomic background (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed, one-tailed test; \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;7.835 x 10\u003csup\u003e\u0026minus;6\u003c/sup\u003e). Our SNP based \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e analysis showed regions surrounding the \u003cem\u003eMc1R\u003c/em\u003e indel had lower divergence than the indel, which underscores the need to analyze the variation at the indel locus directly.\u003c/p\u003e \u003cp\u003eWe also estimated divergence between color morphs in the urban population to corroborate the \u003cem\u003eMc1R\u003c/em\u003e indel locus as the causal locus underlying the coat color polymorphism. \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e between morphs was 0.588 (95% CI 0.523\u0026ndash;0.662) for the \u003cem\u003eMc1R\u003c/em\u003e indel, which was greater than the genomic background (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed, one-tailed test; \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;2.2 x 10\u003csup\u003e\u0026minus;\u0026thinsp;16\u003c/sup\u003e). No other locus in the genome had greater divergence between color morphs, yet our sliding window \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e analysis between morphs revealed the second highest window corresponded to a region on chromosome 12 that contains a single gene \u003cem\u003eSmyd3\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee). In mice, Smyd3 functions in the nucleus, stimulating the transcription of several key regulators involved in cell proliferation, epithelial-mesenchymal transition, and is implicated in cancer development (30, 31). These results support previous findings (21) showing the \u003cem\u003eMc1R\u003c/em\u003e indel is the causal locus for the coat color polymorphism, and presents another region where the connection to coat color expression is not obvious and warrants further investigation.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEstimating the selection coefficient for the coat color allele\u003c/h3\u003e\n\u003cp\u003eWe found much stronger urban-rural divergence at the \u003cem\u003eMc1R\u003c/em\u003e indel locus than the genomic background, pointing to selection playing a significant role in maintaining the urban-rural cline in coat color. To provide insight into the nature of selection in each habitat, we estimated the selection coefficient (\u003cem\u003es\u003c/em\u003e) against the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele in urban and rural environments using three independent approaches.\u003c/p\u003e \u003cp\u003eModel 1 estimated selection analytically from a model of migration-selection balance (32\u0026ndash;34) that assumes the equilibrium frequency difference is determined by the balance between selection removing deleterious alleles and migration introducing deleterious alleles. We used estimates of migration from our demographic model based on WGS data, and we derived contemporary allele frequencies from estimates of the proportion of melanic squirrels in urban and rural environments from standardized field surveys (16). Model 2 compared survival time observations from a translocation experiment (16) to simulations, using Approximate Bayesian Computation (ABC) to estimate the probability of mortality and reproduction, from which we estimated the selection coefficient in urban and rural environments. Model 3 forward simulated allele frequencies using a flexible population model of selection and drift, comprising six parameters that encompassed uncertainty (and prior beliefs) of effective population sizes of both rural and urban populations, generation times, selection strengths, and the ratio of melanics to grays (see Supplemental Methods for full explanation of informative priors). It then used ABC to select the best simulations by comparing observed modern morph frequencies to simulated frequencies.\u003c/p\u003e \u003cp\u003eThe results of all three models show strong selection against the melanic morph in the rural environment, but no selection in the urban environment. The consistency between all three approaches is noteworthy (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and suggests remarkably strong selection in rural forests.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSummary of selection coefficient estimates from all three independent methods: Model 1: migration-selection balance, Model 2: translocation simulations, Model 3: allele frequency simulations.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cem\u003eSource data used to fit model\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003eSelection against melanics relative to gray in urban environment\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eSelection against melanics relative to gray in rural environment\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eModel 1\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eContemporary allele frequencies in urban and rural populations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.008 (95% CI = -0.027 to 0.021)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.086 (95% CI = -0.052 to 0.313)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eModel 2\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAdult survival in urban and rural populations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.000 (95% CI = -0.128 to 0.127)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.234 (95% CI\u0026thinsp;=\u0026thinsp;0.034 to 0.519)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cem\u003eModel 3\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eForward simulation of allele frequencies in urban and rural populations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.019 (95% CI = -0.115 to 0.155)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.109 (95% CI\u0026thinsp;=\u0026thinsp;0.042 to 0.185)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur work provides a detailed examination of the adaptive and demographic processes that contribute to the establishment and maintenance of urban-rural clines in trait variation. We found genomic differentiation between urban and rural populations of gray squirrels was significantly greater at \u003cem\u003eMc1R\u003c/em\u003e than the genomic background, suggesting selection plays a role in maintaining coat color clines between urban and rural environments in cities despite significant historical population bottlenecks. By estimating selection coefficients from three independent datasets using three different methodologies, we found strong support for selection against the melanic morph in rural forests, and support for (near) neutrality at the coat color locus in the urban environment. These findings add to our growing understanding of biological evolution across urbanization gradients, emphasizing the role of rural selection in maintaining urban-rural clines and the potential for cities to function as havens that protect and maintain genetic diversity.\u003c/p\u003e \u003cp\u003eDetecting very recent selective sweeps with classic population genetic approaches can be challenging. Given that we are examining selection on a very recent time scale, and that the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele predates its introduction to cities (35), the allele may have legacy signatures of molecular evolution prior to urbanization. Forward simulations can be helpful in overcoming this challenge because recent changes to allele frequencies occur on the backdrop of standing variation shaped by historical processes. Our two forward simulation models used different input data: one phenotypic survival estimates, the other modern allele frequencies. These models also made very different prior assumptions about the nature of selection. Despite their differences, each model indicated selection against the melanic morph in the rural environment and neutrality for coat color selection in the urban environment (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Estimates of the selection coefficients from an analytical model of migration-selection balance supported the same conclusion. This congruence across three independent approaches \u0026ndash; each with different assumptions and data inputs \u0026ndash; provides compelling evidence that rural populations experience directional selection for the gray morph, whereas coat color appears to be evolving neutrally or nearly so in the city.\u003c/p\u003e \u003cp\u003eThe decline in prevalence of melanics in rural environments coincided with a period of extensive forest loss, agricultural intensification, and severe hunting pressure that accompanied the growth of cities. We suspect the altered structure of sparse and therefore lighter regrown forests relative to denser, darker old growth forests made the melanic morph more visible and susceptible to predation, including and especially hunting by people (17). Meanwhile, squirrels were introduced to cities where firearm use - and by extension, hunting - was increasingly prohibited through municipal ordinances for public safety (34, 35), thereby relaxing a source of selection on coat color that remains important in rural forests (18). However, the lack of a signal of selection on coat color in the city does not render novel environmental change in the city unimportant. Our estimated selection coefficients reflect the \u003cem\u003enet outcome\u003c/em\u003e of multiple possible constraints on fitness. In the urban environment, the fitness benefits melanic squirrels gain from reduced road mortality may offset costs from other processes (e.g., predation), effectively relaxing the total selection on coat color compared to that in rural environments. A more complete understanding of the evolution of coat color along urbanization gradients will require more investigation of the ecological causes of selection in both urban and rural environments, including the potential for direct and indirect selection via correlated traits (e.g., stress response, energy homeostasis, immunity, and resistance to oxidative stress; (36)). Data from multiple cities are also needed to evaluate the generalities of the findings from this study to clarify whether similar ecological conditions in cities and rural environments alike lead to parallel urban-rural clines, or whether convergence is limited by ecological context and nonadaptive processes (15).\u003c/p\u003e \u003cp\u003eIn conclusion, our results suggest the high prevalence of squirrel melanism in some cities can be maintained via relaxation of strong selection against the melanic morph outside the city. The mean selection coefficients in the rural population range based on our models ranged from 0.086 to 0.234, reflecting intense selection relative to values typical in other systems (median \u003cem\u003es\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.082 across \u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;3000 estimates; (37)). Although novel environmental change in cities may play a role in relaxing total selection on coat color, our results challenge the widespread assumption that city environments are the primary source of selection involved in the maintenance of urban-rural clines in trait variation observed across many species.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eSampling and DNA extraction\u003c/h2\u003e \u003cp\u003eOne hundred and ninety-six eastern gray squirrels were collected from urban and rural habitats in and around the general area of Syracuse, New York, USA. Details are provided in Supplementary Methods. Genomic DNA was extracted from a tail tissue or blood as described in (38). Library pools for 46 samples were sent for sequencing at the Yale Center for Genomics, using the NovaSeq 6000 (paired-end 2 x 150 bp reads). DNA from 150 samples were genotyped for the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele using polymerase chain reaction (PCR).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eWhole genome raw data processing and SNP genotyping\u003c/h2\u003e \u003cp\u003eRaw sequences were cleaned and aligned to the \u003cem\u003eS. carolinensis\u003c/em\u003e reference genome (39) (see details in Supplementary Methods). For low and medium coverage sequence data (\u0026lt;\u0026thinsp;20X), using genotype likelihoods is generally recommended over called genotypes (40) given that lower coverage may increase the influence of sequencing or mapping errors. Therefore, we estimated genotype likelihoods using the SAMtools modified MAQ model (41) implemented in ANGSD v.0.938 (25) and used for most of the following analyses (unless otherwise specified). However, certain analyses still require the use of called genotypes; in these cases, we used ANGSD to directly call genotypes. To ensure the robustness of results generated from called genotypes, we reran all analyses performed using genotype likelihoods (aside from the initial linkage disequilibrium estimates) with called genotypes and checked for significant deviations. Additionally, we excluded the sex chromosomes from each of the following analyses to minimize the potential impact of sex differences on results. The full specification of ANGSD parameters used for SNP genotyping, as well as the code for each of the following analyses, can be found at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/agentzero93/syr_squirrel_lcwgs\u003c/span\u003e\u003cspan address=\"https://github.com/agentzero93/syr_squirrel_lcwgs\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eAfter sequencing and read trimming, the 46 lcWGS samples ranged from 80 to 155\u0026nbsp;million paired-end reads (mean\u0026thinsp;=\u0026thinsp;121\u0026nbsp;million). Alignment to the \u003cem\u003eS. carolinensis\u003c/em\u003e reference genome (39) was highly successful for all samples (range 91.3% \u0026minus;\u0026thinsp;94.7%; mean\u0026thinsp;=\u0026thinsp;92.7%), thus all samples were retained for variant calling. A total of 2,180,347,754 sites were identified across 19 autosomes for the 46 samples. Prior to examining population structure between morphs (i.e., melanic vs. gray coat color) and habitats (i.e., urban vs. rural), we filtered SNPs for linkage disequilibrium using ngsLD (42), details in Supplemental Methods. Following pruning based on linkage disequilibrium, the pruned dataset which was utilized for all downstream analyses contained a total of 2,297,744 SNPs (6.5X mean coverage and 1.21% mean missing data).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMC1R-𝚫24 allele genotyping\u003c/h3\u003e\n\u003cp\u003eTo further explore the genetic differentiation associated with the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele, we determined the genotype of all samples. The 46 samples with whole-genome sequencing were assessed visually for the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele in Integrative Genomics Viewer (43) by comparing the cleaned raw sequences to the reference genome. The remaining 150 samples were genotyped for the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele using polymerase chain reaction (PCR), details in Supplementary Methods.\u003c/p\u003e\n\u003ch3\u003ePopulation demography\u003c/h3\u003e\n\u003cp\u003eTo understand population structure and demography of eastern gray squirrels within urban and rural habitats, we first assess if the sample set included closely related individuals (for more information see Supplementary Methods). After confirming individuals were unrelated, we performed an admixture and covariance matrix estimation using PCAngsd (44). The estimated covariance matrix was used as the basis for a principal component analysis (PCA), using the dudi.pca function from the ade4 R package (45). Both the PCA and admixture plots (46) were constructed using R. We tested for isolation-by-distance (IBD) with a Mantel test (47) between a matrix of genetic distances and a matrix of geographic distances using adegenet (48) (see Supplemental Results Figure S8). We calculated \u0026#120645; (nucleotide diversity), and Tajima\u0026rsquo;s \u003cem\u003eD\u003c/em\u003e (49) using ANGSD in 5kb windows with 1kb steps across all autosomes (see Supplemental Results Figure S9).\u003c/p\u003e \u003cp\u003eWe explicitly evaluated several demographic scenarios for urban and rural habitats to explore the demographic history for each habitat. We performed population modeling by inferring demographic parameters from the joint site frequency spectrum (SFS) using coalescent simulations using fastsimcoal2 (23), details in Supplementary Methods.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eGenetic differentiation\u003c/h2\u003e \u003cp\u003eWe estimated population differentiation between habitats (urban vs. rural) and morphs (melanic vs. gray) across the whole genome using two approaches: 1) we estimated genome wide \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e and; 2) genetic differentiation using PCA; and identified outlier loci for subsequent analyses, (details in Supplementary Methods). For the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele we estimated \u003cem\u003eF\u003c/em\u003e\u003csub\u003eST\u003c/sub\u003e using the Weir and Cockerham estimator (29) and a bootstrapping approach, (details in Supplementary Methods).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eSignatures of selective sweeps\u003c/h2\u003e \u003cp\u003eTo identify selective sweeps within outlier regions of interest in both habitat and morph comparisons, we followed the protocol outlined in (28) which uses the ancestral recombination graph (ARG) that describes genealogical relationships and changes to those relationships along the genome due to historical recombination events, (details in Supplementary Methods).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eModel 1: migration-selection balance\u003c/h2\u003e \u003cp\u003eFor the model of migration-selection balance (32\u0026ndash;34), the change in the deleterious allele frequency in a population is given by:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:\\varDelta\\:q=\\frac{-spq[q+h(p-q\\left)\\right]}{1-sq(2hp+q})+mQ-Mq$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eIn this model, there are two components driving allele frequency change. The first is selection (s), and its efficacy is influenced by the relative allele frequencies (p and q) and the dominance coefficient (h). The second is migration, which introduces deleterious alleles based on the immigration rate and the deleterious allele frequency outside the population (mQ). Emigration of deleterious alleles can also occur and is based on the migration rate out of the population and the deleterious allele frequency in the population (Mq). We used observed values of q and Q derived from estimates of the proportion of melanic squirrels in the urban and rural environments in (16). We drew from the posterior distribution of for the proportion melanic from the clinal model and used estimates of m, M, from the demographic model generated using our 46 whole genome samples. We report the results using the dominance coefficient h\u0026thinsp;=\u0026thinsp;1 however, we estimated selection for h\u0026thinsp;=\u0026thinsp;0.7,0.8, and 0.9 as well given that the \u003cem\u003eMc1R\u003c/em\u003e-\u0026#120491;24 allele shows incomplete dominance (Figure S10). Using a bootstrapping approach, (details in Supplementary Methods) we estimated by setting Δq equal to 0 and solved for s.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eModel 2: Translocation simulations\u003c/h2\u003e \u003cp\u003eOur translocation data provided the day of death for 76 adult squirrels, comprising melanic and gray, released into either an urban or rural environment. We generated 10\u0026nbsp;million simulations of 76 agents subjected to a daily death probability sampled from a prior uniform distribution of 0 to 0.4, and used an Approximate Bayesian Computation (ABC) approach to select the best 1000 simulations that most closely matched the death days of the observed data.This provided posterior estimates of the daily death probability for rural melanics (0.0159, 95% CI\u0026thinsp;=\u0026thinsp;0.0102 to 0.0225), urban melanics (0.0060, 95% CI\u0026thinsp;=\u0026thinsp;0.0038 to 0.0086), rural grays (0.0061, 95% CI\u0026thinsp;=\u0026thinsp;0.0038 to 0.0088), and urban grays (0.0060, 95% CI\u0026thinsp;=\u0026thinsp;0.0039 to 0.0084). Kolmogorov-Smirnov tests revealed no difference between urban melanics, rural grays and urban grays, but a significant difference between these three and rural melanics (p\u0026thinsp;\u0026lt;\u0026thinsp;2.2x10\u003csup\u003e\u0026minus;\u0026thinsp;16\u003c/sup\u003e). We then converted the daily death probabilities into reproductive probabilities by assuming reproduction occured at some point between day 5 and day 50 for those that survived. Finally, we converted the reproductive probabilities into selection coefficient estimates (\u003cem\u003es\u003c/em\u003e) as:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/div\u003e\n\u003c/div\u003e\n\u003cp\u003ewhere \u003cem\u003er\u003c/em\u003e is the relative reproductive probabilities (see Supplemental Methods Model 2). These were then converted to positive values for simulations where \u003cem\u003er\u003c/em\u003e\u003csub\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sub\u003e belonged to grays, and negative values where \u003cem\u003er\u003c/em\u003e\u003csub\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sub\u003e belonged to melanics, thus ensuring the interpretation of the selection coefficient is always with respect to grays.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModel 3: Allele frequency simulations.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe forward simulated allele frequencies using a flexible population model of selection and drift (see Figure S5). It modelled a rural population of gray and melanic squirrels starting at 1700 CE (prior to urbanization, with an effective population size sampled from 100 to 100,000 (parameter Nr) using a log uniform distribution to favor smaller population sizes (ensuring uncertainty from the effects of drift were well represented). Modern observed allele frequencies were obtained from published literature (16) (frequency of grays in rural\u0026thinsp;=\u0026thinsp;0.9554, frequency of grays in urban\u0026thinsp;=\u0026thinsp;0.5248), and after generating 1\u0026nbsp;billion simulations under this model, we used ABC to skim the best 5000 simulations, using the sum of the squared differences between simulated and observed modern allele frequencies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u0026nbsp;\u003c/strong\u003eThe authors confirm contribution to the paper as follows: \u003cem\u003estudy conception and design\u003c/em\u003e: M Bonar, AJ Blumenfeld, A Caccone, BJ Cosentino, JP Gibbs, L Campagna; \u003cem\u003edata collection\u003c/em\u003e: NA Fusco, M Bonar, AJ Blumenfeld, BJ Cosentino, JP Gibbs; \u003cem\u003eanalysis and interpretation of results\u003c/em\u003e: M Bonar, AJ Blumenfeld, A Timpson, M G Thomas, A Caccone; \u003cem\u003ewriting of the draft manuscript\u003c/em\u003e: M Bonar; \u003cem\u003emajor comments and edits provided for the draft manuscript\u003c/em\u003e: NA Fusco, L Campagna, A Caccone, BJ Cosentino, JP Gibbs, AJ Blumenfeld, A Timpson, M G Thomas. All authors reviewed the results and approved the final version of the manuscript.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eCompeting Interest Statement:\u003c/strong\u003e The authors declare no competing interest.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSupport for this research was provided by the National Science Foundation (DEB 2017987, DEB 2018140, DEB 2018249). We thank J. Proctor, R. Rich, J. Tooley, and J. Vanek for assistance collecting samples.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eN. B. Grimm, \u003cem\u003eet al.\u003c/em\u003e, Global change and the ecology of cities. \u003cem\u003eScience \u003c/em\u003e\u003cstrong\u003e319\u003c/strong\u003e, 756\u0026ndash;760 (2008).\u003c/li\u003e\n\u003cli\u003eM. Alberti, \u003cem\u003eet al.\u003c/em\u003e, Global urban signatures of phenotypic change in animal and plant populations. \u003cem\u003eProc. Natl. Acad. Sci. \u003c/em\u003e\u003cstrong\u003e114\u003c/strong\u003e, 8951\u0026ndash;8956 (2017).\u003c/li\u003e\n\u003cli\u003eM. T. J. Johnson, J. Munshi-South, Evolution of life in urban environments. \u003cem\u003eScience \u003c/em\u003e\u003cstrong\u003e358\u003c/strong\u003e, eaam8327 (2017).\u003c/li\u003e\n\u003cli\u003eS. W. Breck, S. A. Poessel, P. Mahoney, J. K. Young, The intrepid urban coyote: a comparison of bold and exploratory behavior in coyotes from urban and rural environments. \u003cem\u003eSci. Rep. \u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, 2104 (2019).\u003c/li\u003e\n\u003cli\u003eT. Merckx, \u003cem\u003eet al.\u003c/em\u003e, Urbanization extends flight phenology and leads to local adaptation of seasonal plasticity in \u003cem\u003eLepidoptera\u003c/em\u003e. \u003cem\u003eProc. Natl. Acad. Sci. \u003c/em\u003e\u003cstrong\u003e118\u003c/strong\u003e, e2106006118 (2021).\u003c/li\u003e\n\u003cli\u003eH. Eggenberger, \u003cem\u003eet al.\u003c/em\u003e, Urban bumblebees are smaller and more phenotypically diverse than their rural counterparts. \u003cem\u003eJ. Anim. Ecol. \u003c/em\u003e\u003cstrong\u003e88\u003c/strong\u003e, 1522\u0026ndash;1533 (2019).\u003c/li\u003e\n\u003cli\u003eK. J. Parsons, \u003cem\u003eet al.\u003c/em\u003e, Skull morphology diverges between urban and rural populations of red foxes mirroring patterns of domestication and macroevolution. \u003cem\u003eProc. R. Soc. B Biol. Sci. \u003c/em\u003e\u003cstrong\u003e287\u003c/strong\u003e, 20200763 (2020).\u003c/li\u003e\n\u003cli\u003eK. I. Brans, R. A. Almeida, M. Fajgenblat, Genetic differentiation in pesticide resistance between urban and rural populations of a nontarget freshwater keystone interactor, \u003cem\u003eDaphnia magna\u003c/em\u003e. \u003cem\u003eEvol. Appl. \u003c/em\u003e\u003cstrong\u003e14\u003c/strong\u003e, 2541\u0026ndash;2552 (2021).\u003c/li\u003e\n\u003cli\u003eS. C. Campbell-Staton, \u003cem\u003eet al.\u003c/em\u003e, Parallel selection on thermal physiology facilitates repeated adaptation of city lizards to urban heat islands. \u003cem\u003eNat. Ecol. Evol. \u003c/em\u003e\u003cstrong\u003e4\u003c/strong\u003e, 652\u0026ndash;658 (2020).\u003c/li\u003e\n\u003cli\u003eA. Charmantier, \u003cem\u003eet al.\u003c/em\u003e, How does urbanization affect natural selection? \u003cem\u003eFunct. Ecol.\u003c/em\u003e 1365-2435.14667 (2024). https://doi.org/10.1111/1365-2435.14667.\u003c/li\u003e\n\u003cli\u003eJ. L. Koprowski, K. E. Munroe, A. J. Edelman, \u0026ldquo;Gray not grey: ecology of \u003cem\u003eSciurus carolinensis\u003c/em\u003e in their native range in North America\u0026rdquo; in \u003cem\u003eGrey Squirrels: Ecology and Management of an Invasive Species in Europe\u003c/em\u003e, C. M. Shuttleworth, \u003cem\u003eet al.\u003c/em\u003e, Ed. (European Squirrel Initiative, 2016).\u003c/li\u003e\n\u003cli\u003eD. L. Allen, \u003cem\u003eMichigan fox squirrel management\u003c/em\u003e (Michigan Dept. Of Conservation, 1943).\u003c/li\u003e\n\u003cli\u003eA. W. Schorger, \u0026ldquo;Squirrels in early Wisconsin\u0026rdquo; in \u003cem\u003eTransactions of the Wisconsin Academy of Sciences, Arts, and Letters\u003c/em\u003e, (1949).\u003c/li\u003e\n\u003cli\u003eG. I. Robertson, \u0026ldquo;Distribution of color morphs of \u003cem\u003eSciurus carolinensis\u003c/em\u003e in eastern North America,\u0026rdquo; University of Western Ontario. (1973).\u003c/li\u003e\n\u003cli\u003eB. J. Cosentino, J. P. Gibbs, Parallel evolution of urban\u0026ndash;rural clines in melanism in a widespread mammal. \u003cem\u003eSci. Rep. \u003c/em\u003e\u003cstrong\u003e12\u003c/strong\u003e, 1752 (2022).\u003c/li\u003e\n\u003cli\u003eB. J. Cosentino, J. P. Vanek, J. P. Gibbs, Rural selection drives the evolution of an urban\u0026ndash;rural cline in coat color in gray squirrels. \u003cem\u003eEcol. Evol. \u003c/em\u003e\u003cstrong\u003e13\u003c/strong\u003e, e10544 (2023).\u003c/li\u003e\n\u003cli\u003eJ. Proctor, A. Bryan, B. J. Cosentino, J. P. Gibbs, Crypsis in a polymorphic mammal along an urbanization gradient. \u003cem\u003eUrban Ecosyst. \u003c/em\u003e\u003cstrong\u003e28\u003c/strong\u003e, 94 (2025).\u003c/li\u003e\n\u003cli\u003eR. A. Mccleery, Changes in fox squirrel anti-predator behaviors across the urban\u0026ndash;rural gradient. \u003cem\u003eLandsc. Ecol. \u003c/em\u003e\u003cstrong\u003e24\u003c/strong\u003e, 483\u0026ndash;493 (2009).\u003c/li\u003e\n\u003cli\u003eM. A. Bowers, B. Breland, Foraging of gray squirrels on an urban‐rural gradient: use of the GUD to assess anthropogenic impact. \u003cem\u003eEcol. Appl. \u003c/em\u003e\u003cstrong\u003e6\u003c/strong\u003e, 1135\u0026ndash;1142 (1996).\u003c/li\u003e\n\u003cli\u003eA. F. Parlin, \u003cem\u003eet al.\u003c/em\u003e, Road mortality contributes to the evolution of an urban\u0026ndash;rural cline in squirrel coat color. \u003cem\u003eEvol. Appl. \u003c/em\u003e\u003cstrong\u003e18\u003c/strong\u003e, e70109 (2025).\u003c/li\u003e\n\u003cli\u003eH. McRobie, A. Thomas, J. Kelly, The genetic basis of melanism in the gray squirrel (\u003cem\u003eSciurus carolinensis\u003c/em\u003e). \u003cem\u003eJ. Hered. \u003c/em\u003e\u003cstrong\u003e100\u003c/strong\u003e, 709\u0026ndash;714 (2009).\u003c/li\u003e\n\u003cli\u003eE. Benson, The urbanization of the eastern gray squirrel in the United States. \u003cem\u003eJ. Am. Hist. \u003c/em\u003e\u003cstrong\u003e100\u003c/strong\u003e, 691\u0026ndash;710 (2013).\u003c/li\u003e\n\u003cli\u003eL. Excoffier, \u003cem\u003eet al.\u003c/em\u003e, \u003cem\u003efastsimcoal2\u003c/em\u003e : demographic inference under complex evolutionary scenarios. \u003cem\u003eBioinformatics \u003c/em\u003e\u003cstrong\u003e37\u003c/strong\u003e, 4882\u0026ndash;4885 (2021).\u003c/li\u003e\n\u003cli\u003eN. A. Fusco, \u003cem\u003eet al.\u003c/em\u003e, Population genomic structure of a widespread, urban‐dwelling mammal: The eastern grey squirrel (\u003cem\u003eSciurus carolinensis\u003c/em\u003e). \u003cem\u003eMol. Ecol. \u003c/em\u003e\u003cstrong\u003e33\u003c/strong\u003e, e17230 (2024).\u003c/li\u003e\n\u003cli\u003eT. S. Korneliussen, A. Albrechtsen, R. Nielsen, ANGSD: analysis of next generation sequencing data. \u003cem\u003eBMC Bioinformatics \u003c/em\u003e\u003cstrong\u003e15\u003c/strong\u003e, 356 (2014).\u003c/li\u003e\n\u003cli\u003eR Core Team, R: A language and environment for statistical computing. (2022). Deposited 2022.\u003c/li\u003e\n\u003cli\u003eK. Luu, E. Bazin, M. G. B. Blum, \u003cem\u003epcadapt\u003c/em\u003e: an R package to perform genome scans for selection based on principal component analysis. \u003cem\u003eMol. Ecol. Resour. \u003c/em\u003e\u003cstrong\u003e17\u003c/strong\u003e, 67\u0026ndash;77 (2017).\u003c/li\u003e\n\u003cli\u003eH. A. Hejase, \u003cem\u003eet al.\u003c/em\u003e, Genomic islands of differentiation in a rapid avian radiation have been driven by recent selective sweeps. \u003cem\u003eProc. Natl. Acad. Sci. \u003c/em\u003e\u003cstrong\u003e117\u003c/strong\u003e, 30554\u0026ndash;30565 (2020).\u003c/li\u003e\n\u003cli\u003eB. Weir, C. Clark Cockerham, Estimating F-statistics for the analysis of population structure. \u003cem\u003eEvolution \u003c/em\u003e\u003cstrong\u003e38\u003c/strong\u003e, 1358\u0026ndash;1370 (1984).\u003c/li\u003e\n\u003cli\u003eM. E. Sarris, P. Moulos, A. Haroniti, A. Giakountis, I. Talianidis, Smyd3 is a transcriptional potentiator of multiple cancer-promoting genes and required for liver and colon cancer development. \u003cem\u003eCancer Cell \u003c/em\u003e\u003cstrong\u003e29\u003c/strong\u003e, 354\u0026ndash;366 (2016).\u003c/li\u003e\n\u003cli\u003eD. Yang, \u003cem\u003eet al.\u003c/em\u003e, H3K4 methyltransferase Smyd3 mediates vascular smooth muscle cell proliferation, migration, and neointima formation. \u003cem\u003eArterioscler. Thromb. Vasc. Biol. \u003c/em\u003e\u003cstrong\u003e41\u003c/strong\u003e, 1901\u0026ndash;1914 (2021).\u003c/li\u003e\n\u003cli\u003eJ. B. S. Haldane, A mathematical theory of natural and artificial selection. (Part VI, Isolation.). \u003cem\u003eMath. Proc. Camb. Philos. Soc. \u003c/em\u003e\u003cstrong\u003e26\u003c/strong\u003e, 220\u0026ndash;230 (1930).\u003c/li\u003e\n\u003cli\u003eS. Wright, Evolution in Mendelian populations. \u003cem\u003eGenetics \u003c/em\u003e\u003cstrong\u003e16\u003c/strong\u003e, 97\u0026ndash;159 (1931).\u003c/li\u003e\n\u003cli\u003eH. E. Hoekstra, K. E. Drumm, M. W. Nachman, Ecological genetics of adaptive color polymorphism in pocket mice: geographic variation in selected and neutral genes. \u003cem\u003eEvolution \u003c/em\u003e\u003cstrong\u003e58\u003c/strong\u003e, 1329\u0026ndash;1341 (2004).\u003c/li\u003e\n\u003cli\u003eH. R. McRobie, N. D. Moncrief, N. I. Mundy, Multiple origins of melanism in two species of North American tree squirrel (\u003cem\u003eSciurus\u003c/em\u003e). \u003cem\u003eBMC Evol. Biol. \u003c/em\u003e\u003cstrong\u003e19\u003c/strong\u003e, 140 (2019).\u003c/li\u003e\n\u003cli\u003eA. Roulin, A.-L. Ducrest, Association between melanism, physiology and behaviour: a role for the melanocortin system. \u003cem\u003eEur. J. Pharmacol. \u003c/em\u003e\u003cstrong\u003e660\u003c/strong\u003e, 226\u0026ndash;233 (2011).\u003c/li\u003e\n\u003cli\u003eT. J. Thurman, R. D. H. Barrett, The genetic consequences of selection in natural populations. \u003cem\u003eMol. Ecol. \u003c/em\u003e\u003cstrong\u003e25\u003c/strong\u003e, 1429\u0026ndash;1448 (2016).\u003c/li\u003e\n\u003cli\u003eN. A. Fusco, E. Pehek, J. Munshi‐South, Urbanization reduces gene flow but not genetic diversity of stream salamander populations in the New York City metropolitan area. \u003cem\u003eEvol. Appl. \u003c/em\u003e\u003cstrong\u003e14\u003c/strong\u003e, 99\u0026ndash;116 (2021).\u003c/li\u003e\n\u003cli\u003eD. Mead, \u003cem\u003eet al.\u003c/em\u003e, The genome sequence of the eastern grey squirrel, \u003cem\u003eSciurus carolinensis\u003c/em\u003e Gmelin, 1788. \u003cem\u003eWellcome Open Res. \u003c/em\u003e\u003cstrong\u003e5\u003c/strong\u003e, 27 (2020).\u003c/li\u003e\n\u003cli\u003eR. Nielsen, J. S. Paul, A. Albrechtsen, Y. S. Song, Genotype and SNP calling from next-generation sequencing data. \u003cem\u003eNat. Rev. Genet. \u003c/em\u003e\u003cstrong\u003e12\u003c/strong\u003e, 443\u0026ndash;451 (2011).\u003c/li\u003e\n\u003cli\u003eH. Li, \u003cem\u003eet al.\u003c/em\u003e, The sequence alignment/map format and SAMtools. \u003cem\u003eBioinformatics \u003c/em\u003e\u003cstrong\u003e25\u003c/strong\u003e, 2078\u0026ndash;2079 (2009).\u003c/li\u003e\n\u003cli\u003eE. A. Fox, A. E. Wright, M. Fumagalli, F. G. Vieira, \u003cem\u003engsLD\u003c/em\u003e : evaluating linkage disequilibrium using genotype likelihoods. \u003cem\u003eBioinformatics \u003c/em\u003e\u003cstrong\u003e35\u003c/strong\u003e, 3855\u0026ndash;3856 (2019).\u003c/li\u003e\n\u003cli\u003eJ. T. Robinson, \u003cem\u003eet al.\u003c/em\u003e, Integrative genomics viewer. \u003cem\u003eNat. Biotechnol. \u003c/em\u003e\u003cstrong\u003e29\u003c/strong\u003e, 24\u0026ndash;26 (2011).\u003c/li\u003e\n\u003cli\u003eJ. Meisner, A. Albrechtsen, Inferring Population Structure and Admixture Proportions in Low-Depth NGS Data. \u003cem\u003eGenetics \u003c/em\u003e\u003cstrong\u003e210\u003c/strong\u003e, 719\u0026ndash;731 (2018).\u003c/li\u003e\n\u003cli\u003eS. Dray, A.-B. Dufour, The \u003cstrong\u003eade4\u003c/strong\u003e package: implementing the duality diagram for ecologists. \u003cem\u003eJ. Stat. Softw. \u003c/em\u003e\u003cstrong\u003e22\u003c/strong\u003e (2007).\u003c/li\u003e\n\u003cli\u003eR. M. Francis, pophelper : an R package and web app to analyse and visualize population structure. \u003cem\u003eMol. Ecol. Resour. \u003c/em\u003e\u003cstrong\u003e17\u003c/strong\u003e, 27\u0026ndash;32 (2017).\u003c/li\u003e\n\u003cli\u003eP. E. Smouse, J. C. Long, R. R. Sokal, Multiple regression and correlation extensions of the mantel test of matrix correspondence. \u003cem\u003eSyst. Zool. \u003c/em\u003e\u003cstrong\u003e35\u003c/strong\u003e, 627 (1986).\u003c/li\u003e\n\u003cli\u003eT. Jombart, \u003cem\u003eadegenet\u003c/em\u003e : a R package for the multivariate analysis of genetic markers. \u003cem\u003eBioinformatics \u003c/em\u003e\u003cstrong\u003e24\u003c/strong\u003e, 1403\u0026ndash;1405 (2008).\u003c/li\u003e\n\u003cli\u003eF. Tajima, Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. \u003cem\u003eGenetics \u003c/em\u003e\u003cstrong\u003e595\u003c/strong\u003e, 585\u0026ndash;595 (1989).\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Supplementary Tables","content":"\u003cp\u003eSupplementary Tables are not available with this version\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Yale University","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"selection scan, demographic modelling, selection coefficient, urban evolution, Sciurus","lastPublishedDoi":"10.21203/rs.3.rs-6761695/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6761695/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eUrbanization creates heterogeneous selective landscapes that cause the evolution of urban-rural clines in phenotypic traits. Although cities can introduce novel selection pressures, little attention has been paid to the role of selection outside the city in maintaining urban-rural clines. Here we integrate whole genome sequencing, demographic modeling, and complementary models of selection to test how natural selection in both urban and rural environments shapes the evolution of an urban-rural cline in coat color in eastern gray squirrels (\u003cem\u003eSciurus carolinensis\u003c/em\u003e). Coat color polymorphism in this species, which presents as either gray or melanic, is primarily determined by a 24-bp deletion in the melanocortin-1 receptor gene (\u003cem\u003eMc1R\u003c/em\u003e). Melanic squirrels are often more prevalent in urban environments but rare or absent in rural forests. Whole genome sequencing and demographic modeling revealed substantially greater urban-rural divergence at \u003cem\u003eMc1R \u003c/em\u003ethan the genomic background, suggesting urban-rural clines in melanism are maintained by selection. We applied three separate approaches leveraging demographic and genomic data to estimate the selection coefficient against \u003cem\u003eMc1R \u003c/em\u003ealleles in each habitat, producing a surprising, yet consistent finding: strong selection against the melanic morph in the rural environment and neutrality or near neutrality in the city. Our findings demonstrate that selection outside the city can be sufficient to maintain urban-rural clines, and that urban environments can maintain genetic diversity that would otherwise be lost in rural landscapes. This study provides a rare opportunity to unravel both the spatial dynamics and the selective pressures shaping trait variation in a widespread vertebrate species that thrives across diverse landscapes, highlighting the complex and sometimes protective role of urban landscapes in evolutionary processes.\u003c/p\u003e","manuscriptTitle":"Natural selection beyond city limits","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-29 17:54:16","doi":"10.21203/rs.3.rs-6761695/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f9849d55-2947-411d-9c50-0dd1e2f6c8d0","owner":[],"postedDate":"May 29th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":49127528,"name":"Evolutionary Biology"},{"id":49127529,"name":"Evolutionary Genetics"}],"tags":[],"updatedAt":"2025-05-29T17:54:16+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-29 17:54:16","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6761695","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6761695","identity":"rs-6761695","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}