Genetic influence on the sex ratio of a turtle with temperature-dependent sex determination

Genetic influence on the sex ratio of a turtle with temperature-dependent sex determination | 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 Article Genetic influence on the sex ratio of a turtle with temperature-dependent sex determination Robin Lloyd, Melanie Massey, Joanna Rifkin, Jacqueline Litzgus, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9324153/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract In species with temperature-dependent sex determination (TSD), incubation temperature during embryonic development determines sex. Because sex ratios influence population stability, environmental control of sex ratio makes populations vulnerable to warming, where increases in incubation temperature can result in sex ratio skew. The capacity of populations with TSD to adapt to warmer thermal regimes depends on whether traits related to the thermal sensitivity of sexual outcomes have a genetic basis. Here, we explore the genetic basis of sex in a species with TSD. We use 271 hatchling painted turtles ( Chrysemys picta ) reared under controlled conditions at a temperature that produces both sexes, and we couple these data with 474 wild adult turtles. We used genome-wide association (GWAS) to test for genotype–sex associations and genomic best linear unbiased prediction (GBLUP) analyses to assess the predictive power of genome-wide single-nucleotide polymorphisms. GWAS suggested that no individual loci were strongly associated with sex in hatchlings or adults, and GBLUP analyses revealed limited predictive power for sex in adults and hatchlings. However, heritability was ≈0.29 in adults and ≈0.16 in hatchlings, indicating that additive genetic variance explains a measurable portion of variation in sex among lab-reared and wild turtles. Together, these results indicate that sex in TSD turtles can be influenced by many small-effect genetic variants, with a modest additive genetic effect. Our findings demonstrate that while sex in TSD species is environmentally driven, sex also carries a genetic basis that could support adaptation in the face of changing climate. Biological sciences/Genetics/Genomics/Conservation genomics Biological sciences/Evolution/Evolutionary genetics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Sex ratio has long been recognized as a central driver of population dynamics (Fisher, 1930; Bull, 1983; Bull and Charnov, 1989). Theory predicts that natural populations should exhibit a balanced adult sex ratio, as the genetic contribution of each sex to the next generation must be equal. Deviation from a balanced sex ratio would favour overproduction of the rarer sex, and frequency-dependent selection would ultimately stabilize the sex ratio over time (Fisher, 1930; Charnov and Bull, 1977; Bull and Charnov, 1989). Most organisms exhibit genotypic sex determination (GSD), where sex is determined genetically, often through sex chromosomes or sex-determining genes (Bachtrog et al., 2014; Vandeputte, 2016). Species with GSD typically maintain a balanced sex ratio, or nearly so, from birth to adulthood (Bachtrog et al., 2014; Bókony et al., 2019). However, many organisms have environmental sex determination (ESD), where sex depends on external environmental factors ( e.g ., temperature, social cues, salinity) that act after fertilization, typically during embryonic development (Bachtrog et al., 2014; Bókony et al., 2019). Species with ESD frequently deviate from a 1:1 sex ratio at birth and in adulthood, which challenges Fisher’s (1930) theoretical predictions of balanced sex ratios (Bull and Charnov, 1988; Charnov and Bull, 1989). Over the last several decades, unbalanced sex ratios of ESD species have prompted extensive investigation into the adaptive significance of various forms of ESD, accompanied by growing concerns over the potential impacts of rapid environmental change on population sex ratios (Charnov and Bull, 1977; Janzen and Phillips, 2006; Warner and Shine, 2008; Pen et al., 2010). The most widespread form of ESD is temperature-dependent sex determination (TSD), which is commonly found in reptiles, as well as some fishes and invertebrates (Bull, 1980; Ewert et al., 2004). Under TSD, incubation temperature influences gonadal differentiation into ovaries or testes during the thermosensitive period (TSP), which, in reptiles, typically occurs around the middle third of embryonic development (Yntema, 1979; Bull, 1980; Bull, 1983; Janzen and Paukstis, 1991). Under TSD, constant incubation at an extreme temperature may lead to production of a single sex; for instance, in many TSD turtles, males are produced at cool temperatures and females at warm temperatures (see Lockley and Eizaguirre, 2021). At intermediate temperatures there is a transitional range of temperatures (TRT) where both sexes are produced and sex ratios are mixed. The TRT represents a temperature range where there is a transition between an overproduction of one sex to an overproduction of the other sex, with a pivotal temperature (Tpiv) that produces a 1:1 sex ratio (Mrosovsky and Yntema, 1980; Mrosovsky and Pieau, 1991; Lockley and Eizaguirre, 2021). Variation in Tpiv is influenced by maternal effects (Ewert et al., 2004; Schwanz et al., 2010; Carter et al., 2016; Roush and Rhen, 2018) but also has a genetic basis (Janzen and Paukstis, 1991; Mrosovsky and Pieau, 1991; Janzen, 1992; Ewert et al., 2004), and appears in some cases to have evolved across latitudinal gradients, suggesting adaptation to local conditions (Bull et al., 1982; Ewert et al., 2005; Li et al., 2022). In GSD systems, the bipotential gonad develops into either testes or ovaries depending on the presence of a master sex determining gene such as Dmrt1 in birds or Sry in mammals (Tezak et al., 2020; Thépot, 2021). There is no clear evidence of a master sex determining gene in TSD reptiles, but gene regulation underlies temperature-driven gonadal fate: Dmrt1 and Sox9 mRNA expression increase at male-producing temperatures, whereas aromatase, androgen receptor ( Ar ), and Foxl2 are upregulated at female-producing temperatures (Schroeder et al., 2016; Matsumoto and Crews, 2017; Roush and Rhen, 2018; Thépot, 2021). Given that Tpiv exhibits heritability in TSD species (Janzen, 1992; McGaugh and Janzen, 2011; Refsnider and Janzen, 2016), it is likely that genes regulating hormone expression act as proximate mechanisms through which genetic variation in physiological traits contributes to putatively adaptive responses of Tpiv across geographic gradients. Yet, direct genotype–sex associations remain understudied in TSD reptiles (Roush and Rhen, 2018). Most efforts to link genotype with sex have relied on candidate gene approaches, which have produced mixed results. For example, Schroeder et al. (2016) identified an association between the cold-induced RNA binding protein ( CIRBP ) gene and sex in hatchling common snapping turtles ( Chelydra serpentina ), suggesting that genotype may influence sensitivity to temperature during sex determination. By contrast, Matsumoto and Crews (2017) found no association between polymorphisms in the aromatase ( cyp19a1 ) gene and sex in pond sliders ( Trachemys scripta ), indicating that some key regulators are highly conserved and primarily regulated by temperature rather than genetic variation. A key gap in understanding the genetic basis of TSD is the lack of genome-wide analyses testing for sex-specific associations under controlled incubation conditions, assessing the adaptive potential of physiological thresholds such as Tpiv (Roush and Rhen, 2018; Thépot, 2021), and quantifying heritability of sex (Janzen, 1992; McGaugh and Janzen, 2011; Refsnider and Janzen, 2016). Many TSD species may be vulnerable to the threat of rapid climate change, where a sustained warming will bias sex ratios, potentially resulting in population crashes over the next century (Janzen, 1994b; Schwanz et al., 2020; Tomillo, 2022; Staines et al., 2023). Yet, TSD has persisted for over 200 million years, surviving major climatic upheavals, suggesting TSD is not necessarily maladaptive under climate change (Rage, 1998; McGaugh and Janzen, 2011; Silber et al., 2011). Even though sex ratios in TSD species frequently deviate from a 1:1 sex ratio, Fisherian sex-ratio selection (Fisher, 1930) may become increasingly important under rapid climate shifts for stabilizing sex ratios, with genetic variation playing a critical role in responses to stabilizing selection (McGaugh and Janzen, 2011). Adaptive evolutionary responses of traits associated with sex ratio may arise in heritable thermal physiological traits ( e.g. , Tpiv, TRT, TSP) (Mitchell and Janzen, 2010; McGaugh and Janzen, 2011; Krueger et al., 2025), and/or heritable nest-site choice behaviour (McGaugh et al., 2010; Refsnider and Janzen, 2016). Of these, heritable physiological traits are predicted to be the most effective avenue for TSD species to adapt to rapid climate change (McGaugh and Janzen, 2011; Crowther and Schwanz, 2025). This is because developing embryos have little to no capacity to behaviourally thermoregulate (Telemeco et al., 2016), such that natural selection is expected to act directly on variation in their physiology. Heritable nest-site choice, on the other hand, is constrained by the availability of suitable nesting habitat (McGaugh and Janzen, 2011), making changes in nest-site choice much slower than the pace of climate change, and ultimately of potentially limited adaptive potential except under climate extremes (McGaugh and Janzen, 2011; Topping and Valenzuela, 2021; Crowther and Schwanz, 2025). While genetic variation for physiological traits, especially traits related to the Tpiv and TRT, may hold the adaptive potential for TSD species to persist under climate change, the scope of these genetic influences to mediate sex-ratio bias under climate change is unclear (McGaugh et al., 2010; Roush and Rhen, 2018; Crowther and Schwanz, 2025). Heritability of Tpiv and TRT exists in TSD species (reviewed in Refsnider and Janzen, 2016), but heritable variation is not expressed at extreme incubation temperatures, where the influence of temperature overrides any genetic influence on sex ratios that is otherwise found at intermediate temperatures (Roush and Rhen, 2018; Crowther and Schwanz, 2025). Additive genetic variation for Tpiv under extreme climatic conditions is therefore negligible, and a rapid increase in incubation temperatures under climate change may outpace the potential for an adaptive response of TSD species. A key step in understanding the threat posed by climate change is understanding the extent of genetic variation for sex in wild, unmanipulated TSD populations, especially those with sex ratios that have already undergone a thermally induced bias. Under sustained exposure to extreme incubation temperatures, wild populations with biased sex ratios might be expected to exhibit less genetic variation. Wild adults, shaped by long-term natural selection, provide a snapshot of whether genetic effects on sex persist in the environment, and may offer insight into the adaptive potential of TSD traits under climate change. Yet, in the wild, genetic signals can be confounded by environmental heterogeneity and maternal influences such as yolk steroids, which can also affect sex ratios (Bowden et al., 2000; Roush and Rhen, 2018). In contrast, embryos incubated at the Tpiv undergo sex determination when environmental variance is minimized, allowing genetic effects on sex to be more directly expressed. Here, we compare wild adult turtles with TSD to hatchlings incubated at Tpiv, estimating heritability of sex in the wild and in the lab, and testing whether genetic associations with sex are robust in natural populations or masked by environmental noise . The main objectives of this study are to determine whether adult-wild and Tpiv-reared turtles with TSD show genomic regions associated with sex, whether sex has a detectable additive genetic basis, and how accurately sex can be predicted from genetic variation. Our comparative framework provides a critical step toward evaluating the adaptive potential of TSD species under rapid climate change, investigating specific loci associated with sex, and detection of genetic signal in a complex trait. Materials and Methods Study Site and Population – The present study is part of a long-term study on midland painted turtle ( Chrysemys picta marginata ) initiated in 1978 at Wolf Howl Pond (WHP) and West Rose Lake (WRL), two black spruce ( Picea mariana ) bogs in Algonquin Provincial Park, Ontario, Canada (WHP: 45°34′00″N, 78°41′00″W) (see Figure 1). Both lakes are bisected by a 230 m long, 6–7 m wide decommissioned railway embankment built in the 1890s, now part of a park hiking trail. The embankment provides a warm, sparsely vegetated nesting habitat that has been monitored annually for turtle nesting activity since 1990 (Hughes and Brooks, 2006). Painted turtles are a medium-sized freshwater turtle exhibiting male-female pattern TSD, where cooler incubation temperatures produce males and warmer temperatures produce females (see Lockley and Eizaguirre, 2021). Turtles are captured annually by canoe and dipnet in spring (May) and by hand during the nesting season (June – early July). Individuals are permanently marked with unique notches filed into marginal scutes (Cagle, 1939), and in some cases fitted with Passive Integrated Transponder (PIT) tags. Sex is determined using external morphology: males have elongated foreclaws, a cloacal opening farther from the base of the tail, and smaller body size, while females are larger and lack these traits (Ernst and Lovich, 2009; Moldowan et al., 2016, 2017). Males typically reach sexual maturity at 7-10 years (~90 mm carapace length), and females at 10–16 years (~110 mm) within this population (Samson, 2003; Schwarzkopf and Brooks, 1987) with a lifespan exceeding 60 years for many individuals. At WHP, turtle density is approximately 129 individuals per hectare (Rouleau, 2020), with a strongly female-biased sex ratio of 0.29 males to 1 female (Samson, 2003; Rouleau, 2020; Moldowan et al., 2020). Males frequently sire offspring across multiple clutches, and females are capable of storing sperm for several years (Pearse et al., 2002; Hughes, 2011). Blood sampling of adults for genomic sequencing was conducted from 2018 – 2022, with samples collected from the subcarapacial vein and preserved on Whatman® FTA® cards. 2022 Fieldwork and Incubation Methods – We collected eggs from 133 painted turtle clutches between 1 and 25 June 2022. Once a female had nested, she was identified, and three eggs were randomly selected per clutch (mean clutch size ≈ 6.7 eggs) using a randomizer smartphone app (Random: All Things Generator, version 2.10.16). All eggs were weighed and measured at the Algonquin Wildlife Research Station (AWRS). The three randomly selected eggs were transferred to 20 cm x 10 cm plastic containers, half-filled with moist vermiculite (equal parts water and vermiculite), with puncture holes in the lid for airflow. Containers were placed in one of two IN30 Torrey Pines Echo Therm incubators at 27.3°C (the Tpiv for this population, Schwarzkopf and Brooks, 1985, 1987), for approximately three weeks, until the end of the nesting season. Eggs were moved weekly within and between the two incubators to disrupt thermal gradients. To minimize water loss, we compared the current weight of each container to its initial weight every 3-4 days, replenishing evaporated water to each container. After the nesting season, all eggs were transferred to the University of Toronto in an automobile over three hours, and upon arrival at the University, eggs were immediately placed into two Memmert Natural Convection incubators, under the same protocol as used at the AWRS until hatching. Upon hatching, each hatchling was humanely euthanized using an injection of 10 mg/ml Alfaxalone, and 1-2 minutes after the injection, hatchlings were decapitated and immediately pithed. A blood sample was taken from each hatchling for genetic analysis using Whatman® FTA® cards. The sex of the hatchlings was determined via dissection under an Olympus SZX16 dissection microscope. Sex determination was conducted three times by two researchers. The first determination was performed by R.B.L and occurred during the dissection of each hatchling. The second and third determinations were made by R.B.L. and M.D.M. using photos of the Dissected hatchlings. Sex determination was based on the morphology of the gonads: male gonads were small, short, and rounded, while female gonads were long and thin (Yntema and Mrosovsky, 1980; Wyneken et al., 2007). Filtering of SNP Data – Blood samples collected from wild adult turtles and experimentally incubated hatchlings were genotyped for single nucleotide polymorphisms (SNPs) at Diversity Arrays Technology (DarTSeq, Canberra, Australia), which targets hypomethylated, low-copy genomic regions via complexity reduction methodology (see Melville et al., 2017). The DarTSeq platform used SNP loci that were called and aligned to a western painted turtle ( Chrysemys picta belli ) reference genome (GCF_00241765.4) to ensure consistent marker positioning across individuals (Shaffer et al., 2013). We filtered raw SNP data (82,522 loci) from DarTSeq to improve data quality and identify informative SNPs using the R package dartR (v4.3.2; Gruber et al., 2018; R Core Team, 2018). Following the SNP filtering framework described by Gruber et al. (2019), we first removed monomorphic loci, which lack allelic variation and thus provide no information for genetic analyses. We next retained only SNPs with a repeatability ≥ 0.95 to ensure consistent detection across replicates and improve genotyping reliability. We further excluded secondary loci—multiple SNPs from the same sequence fragment—to eliminate redundant information. We then applied a call rate threshold of 0.95, removing loci with more than 5% missing data (Anderson et al., 2010). SNPs with minor allele frequency < 0.01 were removed, as rare alleles offer limited statistical power and may introduce noise (reference). We also removed loci that significantly deviated from Hardy–Weinberg equilibrium (𝛼 < 0.0001), where the p-value reflects the likelihood that deviations in genotype frequencies arise by chance. This stringent threshold is commonly used to eliminate SNPs potentially affected by genotyping errors, thereby improving data quality for downstream analyses (Chen et al., 2017; Ropp et al., 2023). Finally, we filtered out individual turtles with a call rate below 0.95, meaning those missing more than 5% of SNP genotype calls, to ensure that only samples with high genotyping completeness were retained for analysis (Anderson et al., 2010). Approximately 0.63% of SNP genotype calls were missing in the filtered wild adult turtle dataset and 0.65% in the filtered experimental hatchling dataset. To replace missing SNP genotype calls, the most common allele (mode) for each SNP was imputed using the Dplyr package in R, as the genomic analyses required complete genotype data with no missing values (Yarberry, 2021; Appadurai et al., 2023). Genome-wide Association Study (GWAS) – We conducted two separate genome-wide association studies (GWAS) to identify SNPs associated with sex: one for wild adult turtles and one for experimental hatchlings. We separated these groups because adults incubated naturally and were likely subject to post-hatching selection, while hatchlings were reared under controlled, constant conditions at the Tpiv. We coded the filtered SNP data as 0, 1, or 2, representing the number of alternate alleles per locus. The phenotypes were binary-coded as 1 (male) and 0 (female). To control for relatedness and population structure, we used the GMMAT package in R (Chen et al., 2019) to fit a logistic mixed null model of the form: logit(P(sex = 1)) = β₀ + u , where β₀ is the fixed intercept and u is a random effect accounting for kinship among individuals. Relatedness was modeled using a genomic relationship matrix (GRM) constructed following VanRaden (2008) where genotype values were centered by allele frequency and scaled by the total expected heterozygosity across loci to quantify pairwise genomic similarity among individuals. This GRM or kinship matrix was used to model the random genetic effect in the mixed model. The logistic mixed null model was first fit without SNP effects, and each SNP was then tested individually using a score test via the lm.score function in R to assess its association with sex while accounting for relatedness (Rao, 1948; Chen et al., 2016). We then merged P -value results with SNP metadata (chromosome number and genomic position) provided by Diversity Arrays Technology for genomic mapping. For visualization, we then used the qqman (Turner, 2014) and ggplot2 packages (Wickham, 2011) to generate Manhattan and QQ plots to showcase GWAS results. To evaluate potential test statistic inflation, we calculated the genomic inflation factor ( λ ), defined as: λ = median(χ²_observed) / 0.456, where 0.456 is the expected median of a χ² distribution with 1 degree of freedom under the null hypothesis (Yang et al., 2011). The λ value was embedded in each QQ plot to assess the adequacy of population structure correction. Genomic Prediction via Genomic Best Linear Unbiased Prediction (GBLUP) – Genomic prediction was used to assess the extent to which genome-wide SNP variation explains sex in our painted turtle population. We implemented genomic prediction using the BGLR package in R (Pérez and Campos, 2014), applying a Reproducing Kernel Hilbert Space (RKHS) model with a linear genomic kernel, which is equivalent to a standard GBLUP model that uses a GRM to account for additive genetic effects (VanRaden, 2008; Gianola and Kaam, 2008; Wang et al., 2018). The GRM was constructed using the same VanRaden (2008) approach as in the GWAS analyses and used as the genomic kernel in the GBLUP models to account for genome-wide relatedness among individuals. Our GBLUP model was defined as y = μ + u + ε , where y is the vector of binary sex phenotypes (1 = male, 0 = female), μ is the overall mean, u is the vector of additive genetic effects assumed to follow a normal distribution u ~ N(0, Gσ²ᵤ) , and ε is the residual error, assumed to follow ε ~ N(0, Iσ²ₑ) (VanRaden, 2008; Wang et al., 2018; Souza et al., 2024). Bayesian estimation of model parameters was conducted using Gibbs sampling with 3,000 iterations and a 1,000-iteration burn-in (Pérez and Campos, 2014). To evaluate model performance while accounting for sex ratio imbalance, we conducted 500 bootstrap iterations separately for adults and hatchlings, where sampling was stratified to preserve the original sex ratio of the sample. In each iteration, 85% of individuals were randomly selected to form the training set, which was used to fit the prediction model and construct the genomic kernel. The remaining 15% of individuals comprised the test set, which was completely held out during model training. Sex for test individuals was predicted based on their genomic similarity to the training individuals. Model performance was evaluated using four metrics: (1) accuracy, defined as the proportion of correct classifications across males and females (Dekkers and Cheng, 2021); (2) the area under the receiver operating characteristic curve (AUC ) , which measures how well the model distinguishes between males and females across all possible probability thresholds used for classification; values near 0.5 indicate random prediction, while values near 1.0 indicate perfect discrimination (Wray et al., 2010); (3) F1 score, the harmonic mean of precision (the proportion of predicted males that are truly male) divided by recall (the proportion of actual males correctly predicted), which balances false positives and false negatives and is especially informative when sex classes are imbalanced (Saito and Rehmsmeier, 2015); and (4) narrow-sense heritability ( h² ), which quantifies the proportion of phenotypic variance attributable to additive genetic effects (Falconer, 1996; Holland et al., 2003; Srivastava et al., 2023). We estimated narrow-sense heritability on the observed phenotypic scale for each GBLUP model iteration within each bootstrap replicate by calculating h² = Var(u) / [Var(u) + Var(ε)] , where Var(u) is the variance of genomic breeding values and Var(ε) is the residual variance estimated by the GBLUP model (Holland et al., 2003; VanRaden, 2008; Souza et al., 2024). To quantify uncertainty, we calculated 95% confidence intervals for each performance metric and heritability estimate using a nonparametric percentile bootstrap approach based on the empirical distribution across the 500 bootstrap replicates. Results A total of 399 eggs were collected from the study population. Of these, 327 (82%) hatched successfully after incubation at the University of Toronto. Eighteen hatchlings were excluded because sex could not be determined, either due to gonadal tearing during dissection or because the gonads were too underdeveloped to classify. Of the remaining 309 hatchlings, 13 were not confidently sexed from photographs, leaving 296 individuals with confirmed sex for genetic analysis. An additional 25 individuals were removed because of incomplete or failed genotyping, resulting in a final hatchling dataset of 271 turtles (111 females, 160 males) representing 125 distinct mothers. Separately, genotyping was also conducted on wild-caught adult turtles sampled in 2022 and from previous years. After SNP filtering and quality control, the final dataset included 839 individuals and 10,423 SNPs comprised of the 271 hatchlings and 568 wild-caught turtles. Among the wild-caught individuals, sex and genotyping data were successfully obtained for 474 adults (389 females, 85 males) including all mothers of the incubated hatchlings, which were used in downstream genomic analyses. Separate GWAS analyses were conducted for adult and hatchling painted turtles to identify SNPs associated with sex. In both datasets, SNPs were distributed across all 25 chromosomes, and no loci reached conventional genome-wide significance thresholds (Figures 2 and 3). QQ plots indicated minimal inflation in test statistics, with λ = 1.017 for adults and 1.021 for hatchlings (Figure 4a,b), suggesting appropriate control for population structure and relatedness. To assess the predictive power of genome-wide SNPs for sex determination, GBLUP models were applied to both adult and hatchling datasets using 10,423 filtered SNPs across 500 bootstrap replicates. Prediction accuracy was higher in adults than in hatchlings. In adults, the mean prediction accuracy (95% CI) was 0.757 (0.637–0.855), with an average AUC of 0.661 (0.468–0.805) (Figure 5a,b). The F1 score was relatively low at 0.351 (0.135–0.561), reflecting a trade-off between precision and recall, which was likely influenced by the imbalanced sex ratio in the adult sample (Figure 5c). The estimated narrow-sense heritability of sex in adults was 0.29 (0.203–0.391), suggesting that roughly one-third of the variance in sex classification could be attributed to additive genetic effects in the wild population (Figure 5d). The hatchling dataset featured lower predictive performance. The mean prediction accuracy was 0.571 (0.426–0.707), and the AUC was 0.568 (0.452–0.704) (Figure 5a,b). Compared to adults, the F1 score was substantially (but not significantly) greater in hatchlings at 0.662 (0.521–0.779), likely reflecting the more balanced sex ratio in this group compared to the wild adults and improved recall for the minority class (Figure 5c). The estimated heritability of sex in hatchlings was 0.16 (0.114–0.237), slightly but not significantly less than that of the adult wild population (Figure 5d). Discussion Adaptive responses of TSD species to warmer incubation conditions will depend on the microevolution of traits underlying sex determination and the rate at which these traits can evolve relative to environmental change (McGaugh and Janzen, 2011; Refsnider and Janzen, 2016; Topping and Valenzuela, 2021; Krueger et al., 2025). Here we present, to our knowledge, the first within-population comparison of genotype–sex associations between wild adults and hatchlings incubated at the Tpiv in a TSD species. In adults, we detected modest heritability of sex, whereas hatchlings exhibited a slightly (but not significantly) lower heritability, and no single SNP locus was a strong predictor of sex for hatchlings or adults. These results are consistent with a polygenic basis for sex determination (Roush and Rhen, 2018), where sex is influenced by the cumulative effects of many small-effect loci, with no evidence for a significant difference between wild adults and hatchlings incubated at the Tpiv. In the present study, controlled incubation of hatchlings at the Tpiv should have minimized environmental variance and enhanced genetic signal for sex, but we observed no such pattern (Figure 5). One explanation for why hatchlings did not exhibit greater heritability for sex than adults is that genotype-sex associations in wild adults are the outcome of both post-hatching survival and developmental conditions of the embryo. Post-hatching filtering through ontogeny reflects viability selection (Hadfield, 2008), where mortality after hatching disproportionately removes individuals whose genotype–sex combinations are maladaptive, consequently increasing the observed genetic signal by virtue of better-aligning genotype with sex (McGaugh et al., 2011; Mittell et al., 2025). Sex-linked differences in embryonic thermal tolerance (Tomillo, 2022), hormone expression (Crews et al., 1989; Rhen et al., 2007), or metabolic requirements (O’Steen and Janzen, 1999) could drive such filtering. In addition, some of the wild adults may be siblings that developed in the same nesting environment, which can align similar genotypes with sex even when no genetic variation for sex was expressed, giving the appearance of stronger heritable effects when causal loci are weak (McGaugh et al., 2010). While the GRMs in our models account for genomic similarities, they do not explicitly model maternal effects and shared environmental conditions that could align genotype and incubation environment across generations, potentially leading to modest inflation of heritability estimates in wild adults. By contrast, incubation of embryos precisely at the Tpiv reduces between-clutch variation in incubation temperature (Rhen and Lang, 1995), standardizing developmental conditions around the transitional range of temperature, such that sexual outcomes would be most sensitive to maternal effects (e.g., yolk hormones) and genotype (Ewert et al., 2004; Roush and Rhen, 2018) than in wild incubated nests. The absence of viability selection (or weaker viability selection under modest embryo mortality), maternal nest-site choice, and site fidelity in this controlled laboratory setting may reduce genotype–sex correlation, leading to uninflated heritability estimates. Finally, methodological constraints may have also dampened estimates of heritability for hatchlings, as hatchling turtles can be challenging to sex based on gonadal examination (Yntema and Mrosovsky, 1980), leading to mistaken sex classification that dampens estimates of heritability for sex. We did not detect significant SNPs for sex in our GWAS, which is consistent with findings in other TSD systems. While our sample sizes were among the largest yet reported for a reptile GWAS, they remain modest by GWAS standards, where thousands and sometimes millions of samples may be required to detect loci of small effect (Santure and Garant, 2018; Uffelmann et al., 2021). Consequently, our study likely had sufficient power to rule out major effect loci (Santure and Garant, 2018), but probably had limited ability to detect the small effect variants expected under a polygenic architecture. In a study similar to our own, Chow et al. (2019) reported up to 30 loci that appeared to differ between males and females in a TSD loggerhead sea turtles ( Caretta caretta ), suggesting possible sex-specific genotypes. However, their small sample size (n = 45) and low sequencing coverage meant that none reached genome-wide significance, limiting confidence in these associations. Similarly, candidate-gene approaches have often failed to detect robust associations between genotype and sex. For instance, Matsumoto and Crews (2017) found no link between aromatase ( cyp19a1 ) polymorphisms and sex found in pond sliders, whereas CIRBP variants in the common snapping turtle showed associations with sex only under specific incubation conditions (Schroeder et al., 2016). Both patterns are biologically plausible, as genetic variation that influences sex may lie not within coding regions of known sex-related genes, but rather in upstream regulatory elements that influence temperature sensitivity of these genes (Schroeder et al., 2016; Matsumoto and Crews, 2017). Both genome-wide and candidate-gene approaches converge on the conclusion that genetic influences appear to be distributed across many loci of small effect (Santure and Garant, 2018; Chow et al., 2019), with outcomes shaped by interactions between genotype and environment (Roush and Rhen, 2018). Future progress in identifying specific sex-associated loci will likely require large scale GWAS designs involving thousands of individuals, and may necessitate multiple populations to achieve these large sample sizes, where statistical power is sufficient to detect subtle effects and to test whether associations are consistent or population-specific (Santure and Garant, 2018). Our GBLUP analyses revealed limited predictive power for sex. Prediction accuracy in adults was significantly greater than random expectation (≈0.76), but AUC (≈0.66) and F1 score (≈0.35) were not different than random and suggest limited ability to predict sex. High prediction accuracy and low F1 score in adults likely occurred because of the strong female bias in the wild adults, which reduced recall of males and depressed the harmonic mean of precision and recall. While our heritability data suggest a genetic signal is present, patterns observed in our GBLUP analyses suggest that prediction accuracy may be inflated in the adult dataset. In hatchlings, prediction accuracy (≈0.57) and AUC (≈0.57) were not different from random expectations, although the F1 score (≈0.66) was greater than 0.5 and suggested fair discriminatory ability. Indeed, hatchlings had a more balanced sex ratio than adults, which raised the F1 score despite low overall prediction accuracy. The arc of these GBLUP results suggests, at best, that additive genomic variance contributes very weakly to sex outcomes. For comparison, prediction of binary traits with strong genetic control can reach much higher values than what we observed in turtles. In domestic dogs ( Canis lupus familiaris ), the eye disease distichiasis has been predicted with AUC values close to 0.90, showing strong separation between affected and unaffected animals (Thorsrud et al., 2025). Root vigor in sugar beet ( Beta vulgaris ) is controlled by only a few loci and has been predicted with AUC values near 0.98, essentially perfect classification (Biscarini, 2014). However, similar to our findings, sex in European seabass ( Dicentrarchus labrax ), a species with polygenic, temperature-driven sex determination, was predicted with only about 67% accuracy using SNP data (Palaiokostas et al., 2015). Thus, heritability estimates and some features of our GBLUP analyses suggest sex in painted turtles carries a detectable genetic signal, but for GBLUP analyses, prediction accuracy is similar to systems where the environment interacts with polygenic control. Long-standing theory has argued that if traits such as Tpiv are heritable, they can respond to selection imposed by changing climates, including sex-ratio selection (Janzen, 1992; Schwanz et al., 2016; Refsnider and Janzen, 2016). Empirical work among related clutches and lineages supports this prediction, documenting heritable differences in sex-ratio reaction norms among clutches and lineages (Rhen and Lang, 1998; McGaugh and Janzen, 2011; Krueger and Janzen, 2023; Krueger et al., 2025). Indeed, the present study estimated heritability of sex at ≈0.29 in adults and ≈0.16 in hatchlings, values similar to the effective narrow-sense heritability of ≈0.13 for Tpiv of painted turtles in Illinois, USA (McGaugh et al. 2011). Yet, the expression of genetic variation for sex depends on whether nest temperatures fall within the TRT. Indeed, heritability estimates are often high in laboratory studies based on among-clutch (full-sibling) variance under constant incubation conditions near the Tpiv, but substantially lower when expressed at the population level under natural, fluctuating nest temperatures (McGaugh and Janzen, 2011, the present study). Family-level studies showing heritable differences in reaction norms (Rhen and Lang, 1998; McGaugh and Janzen, 2011; Krueger and Janzen, 2023), combined with our SNP-based estimates of the additive genetic variance underlying sex, nevertheless indicate that selection has something to act on. Still, several constraints limit how quickly populations may respond. For instance, much of the additive genetic variance underlying sex determination remains hidden under extreme incubation regimes. Biased adult sex ratios commonly found in TSD systems (Bókony et al. 2019) further reduce the amount of usable genetic variation by lowering effective population size and accelerating genetic drift (Refsnider and Janzen, 2016; Schwanz et al., 2020). Furthermore, the lengthy generation times exhibited by most long-lived TSD species (e.g., Sabath et al. 2016) dampen the pace of any evolutionary response to climate change. In addition, while nest-site choice has been proposed as an alternative heritable pathway for buffering sex ratios, its contribution is constrained by habitat availability and ecological context, making it a slower and less reliable avenue for adaptation than physiological traits related to Tpiv and TRT (Janzen, 1994a; Topping and Valenzuela, 2021; Crowther and Schwanz, 2025). Overall, our results indicate that heritability of traits linked to sex is modest and probably sufficient to support gradual shifts in traits linked to sexual outcomes, but it is unclear whether the pace of adaptation in these traits will be sufficient to support adaptation to rapid climate warming. From a conservation perspective, our results reinforce concerns about the vulnerability of TSD species under rapid climate change. Persistent female-biased sex ratios are already observed in our Algonquin Park population at the northern-most limit of the painted turtle’s geographic range (Moldowan et al., 2020), although the cause of biased sex ratios is likely due to a warm, anthropogenic nesting environment, rather than long-term sustained warming of the environment. As temperatures continue to rise, sex-ratio bias will likely intensify, potentially pushing populations toward demographic collapse if adaptive responses cannot keep pace (Staines et al., 2023). Although the presence of additive genetic variance provides some scope for microevolution, the modest effect sizes we observed suggest that evolutionary rescue is likely to be gradual. This potential mismatch between the rate of environmental change and the strength of genetic response underscores the importance of conservation strategies that buffer populations against skewed sex ratios (Staines et al., 2023), whether through habitat protection (Rhodin et al., 2018), assisted nest management (Wnek et al., 2013), or long-term maintenance of genetic diversity (Velo-Antón et al., 2011). Sustaining population connectivity and preserving genetic diversity may be especially important for facilitating adaptive responses under intensifying climate stress (Velo-Antón et al., 2011; Rhodin et al., 2018). Preserving diversity not only safeguards long-term adaptive capacity but also strengthens the chance of an effective response within the short evolutionary window imposed by rapid climate warming (Janzen, 1994b; McGaugh and Janzen, 2011; Refsnider and Janzen, 2016; Crowther and Schwanz, 2025). Declarations Acknowledgements We thank the EEB Department at the University of Toronto and the Algonquin Wildlife Research Station for providing accommodations and resources during fieldwork. We would like to thank the Congdon-Dickson Turtle Ecology Fund, which funded a generous portion of our genetics work for this project. We are especially grateful to Asher Cutter, Jacqueline Sztepanacz, and John Stinchcombe for their intellectual input during project design. Jessica Leivesley and Mariel Terebiznik also shared valuable expertise for SNP filtering in the early stages of this project and training in identifying sex in hatchling turtles. We would like to thank the thoughtful feedback of Rebecca Schalkowski and Phillip Pearson on an early draft of this manuscript. In addition, we greatly appreciate the field assistance of Lilian Chan and Claire Voss during the turtle nesting season at Algonquin Provincial Park. We thank Diversity Arrays Technology and their team for conducting the genotyping of our blood samples for this project. Lastly, we acknowledge AI-assisted tools (ChatGPT, OpenAI) that were used to assist with debugging, troubleshooting, and organizing code in R software. Funding was provided by NSERC Discovery grants to J.D.L. and N.R. Statement of Authorship RL and NR developed hypothesis and experimental design. RL conducted the fieldwork and laboratory experiment. RL performed the initial sex determinations for specimens with a secondary sex determination opinion from MM. RL performed the data management and statistical analyses with suggestions from JR. NR & JL funded the study and co-manage the painted turtle long-term data collection at the Algonquin Wildlife Research Station. RL and NR led the writing of the manuscript with contributions by all authors. Re search Involving Animals and their Data or Biological Material Animal collecting and use was authorized a Wildlife Scientific Collectors Authorization #1100425 issued by the Ministry of Northern Development, Mines, Natural Resources and Forestry, and an Animal Use Protocol #20011948 approved and issued by the University of Toronto Local Animal Care Committee. References Anderson CA, Pettersson FH, Clarke GM et al. (2010) Data quality control in genetic case-control association studies. Nat Protoc 5:1564–1573. https://doi.org/10.1038/nprot.2010.116 Appadurai V, Bybjerg-Grauholm J, Krebs MD, Rosengren A, Buil A, Ingason A et al. (2023) Accuracy of haplotype estimation and whole genome imputation affects complex trait analyses in complex biobanks. Commun Biol 6:101 Bachtrog D, Mank JE, Peichel CL, Kirkpatrick M, Otto SP, Ashman TL et al. (2014) Sex determination: why so many ways of doing it? PLoS Biol 12:e1001899 Biscarini F, Stevanato P, Broccanello C, Stella A, Saccomani M (2014) Genome-enabled predictions for binomial traits in sugar beet populations. BMC Genet 15:87 Bókony V, Milne G, Pipoly I, Székely T, Liker A (2019) Sex ratios and bimaturism differ between temperature-dependent and genetic sex-determination systems in reptiles. BMC Evol Biol 19:57 Bowden RM, Ewert MA, Nelson CE (2000) Environmental sex determination in a reptile varies seasonally and with yolk hormones. Proc Biol Sci 267:1745–1749 Bull JJ (1980) Sex determination in reptiles. Q Rev Biol 55:3–21 Bull JJ, Charnov E (1988) How fundamental are Fisherian sex ratios? [still needs journal, volume, and page range] Bull JJ, Charnov EL (1989) Enigmatic reptilian sex ratios. Evolution 43:1561–1566 Bull JJ, Vogt RC, McCoy CJ (1982) Sex determining temperatures in turtles: a geographic comparison. Evolution 36:326–332 Carter AW, Bowden RM, Paitz RT (2017) Seasonal shifts in sex ratios are mediated by maternal effects and fluctuating incubation temperatures. Funct Ecol 31:876–884 Charnov EL, Bull J (1977) When is sex environmentally determined? Nature 266:828–830 Chen B, Cole JW, Grond-Ginsbach C (2017) Departure from Hardy Weinberg equilibrium and genotyping error. Front Genet 8:300354 Chen H, Wang C, Conomos MP, Stilp AM, Li Z, Sofer T et al. (2016) Control for population structure and relatedness for binary traits in genetic association studies via logistic mixed models. Am J Hum Genet 98:653–666 Chow JC, Anderson PE, Shedlock AM (2019) Sea turtle population genomic discovery: global and locus-specific signatures of polymorphism, selection, and adaptive potential. Genome Biol Evol 11:2797–2806 Crews D, Wibbels T, Gutzke WHN (1989) Action of sex steroid hormones on temperature-induced sex determination in the snapping turtle (Chelydra serpentina). Gen Comp Endocrinol 76:159–166 Crowther C, Schwanz LE (2025) Behavioural vs. physiological adaptation: which contributes more to the evolution of complex traits in a warming climate? J Evol Biol 38:467–480 Dekkers JC, Su H, Cheng J (2021) Predicting the accuracy of genomic predictions. Genet Sel Evol 53:1–23 Dempster ER, Lerner IM (1950) Heritability of threshold characters. Genetics 35:212 Ewert MA, Etchberger CR, Nelson CE (2004) Turtle sex-determining modes and TSD patterns, and some TSD pattern correlates. In: Valenzuela N, Lance VA (eds) Temperature-dependent sex determination in vertebrates. Smithsonian Books, Washington, DC, pp 21–32 Ewert MA, Lang JW, Nelson CE (2005) Geographic variation in the pattern of temperature-dependent sex determination in the American snapping turtle (Chelydra serpentina). J Zool 265:81–95 Falconer DS (1996) Introduction to quantitative genetics. Pearson Education India, New Delhi Fisher RA (1930) The genetical theory of natural selection. Clarendon Press, Oxford Gianola D, Van Kaam JB (2008) Reproducing kernel Hilbert spaces regression methods for genomic assisted prediction of quantitative traits. Genetics 178:2289–2303 Gruber B, Unmack P, Berry O, Georges A (2019) Introduction to dartR. User Manual 51:1–41 Hadfield JD (2008) Estimating evolutionary parameters when viability selection is operating. Proc R Soc B 275:723–734 Holland JB, Nyquist WE, Cervantes-Martínez CT, Janick J (2003) Estimating and interpreting heritability for plant breeding: an update. Plant Breed Rev 22:9–112 Hughes EJ, Brooks RJ (2006) The good mother: does nest-site selection constitute parental investment in turtles? Can J Zool 84:1545–1554 Hughes EJ (2011) The effect of sex ratio on male reproductive success in painted turtles, Chrysemys picta. Dissertation, University of Guelph Janzen FJ, Paukstis GL (1991) Environmental sex determination in reptiles: ecology, evolution, and experimental design. Q Rev Biol 66:149–179 Janzen FJ (1994a) Vegetational cover predicts the sex ratio of hatchling turtles in natural nests. Ecology 75:1593–1599 Janzen FJ (1994b) Climate change and temperature-dependent sex determination in reptiles. Proc Natl Acad Sci USA 91:7487–7490 Janzen FJ (1992) Heritable variation for sex ratio under environmental sex determination in the common snapping turtle (Chelydra serpentina). Genetics 131:155–161 Janzen FJ, Phillips PC (2006) Exploring the evolution of environmental sex determination, especially in reptiles. J Evol Biol 19:1775–1784 Krueger CJ, Janzen FJ (2023) On the origin of patterns of temperature-dependent sex determination. Evolution 77:1091–1100 Krueger CJ, Girondot M, Janzen FJ (2025) The tortoise and the air: climate shapes sex-ratio reaction norm variation in turtles. Evolution qpaf126 Li S, Xu Z, Luo L, Ping J, Zhou H, Xie L, Zhang Y (2022) Latitudinal variation in the pattern of temperature-dependent sex determination in the Japanese gecko, Gekko japonicus. Animals 12:942 Lockley EC, Eizaguirre C (2021) Effects of global warming on species with temperature-dependent sex determination: bridging the gap between empirical research and management. Evol Appl 14:2361–2377 McGaugh SE, Schwanz LE, Bowden RM, Gonzalez JE, Janzen FJ (2010) Inheritance of nesting behaviour across natural environmental variation in a turtle with temperature-dependent sex determination. Proc R Soc B 277:1219–1226 Matsumoto Y, Crews D (2017) Genetic polymorphisms in aromatase (cyp19a1) are not associated with gonadal phenotypes in red-eared slider turtle hatchlings developed at a pivotal temperature. Sex Dev 11:151–160. https://doi.org/10.1159/000471940 McGaugh SE, Janzen FJ (2011) Effective heritability of targets of sex-ratio selection under environmental sex determination. J Evol Biol 24:784–794 McGaugh S, Bowden R, Kuo CH, Janzen F (2011) Field-measured heritability of the threshold for sex determination in a turtle with temperature-dependent sex determination. Evol Ecol Res 13:75–90 Melville J, Haines ML, Boysen K, Hodkinson L, Kilian A, Smith Date KL et al. (2017) Identifying hybridization and admixture using SNPs: application of the DArTseq platform in phylogeographic research on vertebrates. R Soc Open Sci 4:161061 Mitchell NJ, Janzen FJ (2010) Temperature-dependent sex determination and contemporary climate change. Sex Dev 4:129–140 Mittell EA, Pemberton JM, Kruuk LEB, Morrissey MB (2025) Unmeasured prior viability selection resolves the paradox of stasis for body size in wild Soay sheep. Proc Natl Acad Sci USA 122:e2513969122 Mrosovsky N, Pieau C (1991) Transitional range of temperature, pivotal temperatures and thermosensitive stages for sex determination in reptiles. Amphibia-Reptilia 12:169–179 Mrosovsky N, Yntema CL (1980) Temperature dependence of sexual differentiation in sea turtles: implications for conservation practices. Biol Conserv 18:271–280 O’Steen S, Janzen FJ (1999) Embryonic temperature affects metabolic compensation and thyroid hormones in hatchling snapping turtles. Physiol Biochem Zool 72:520–533 Palaiokostas C, Bekaert M, Taggart JB, Gharbi K, McAndrew BJ, Chatain B et al. (2015) A new SNP-based vision of the genetics of sex determination in European sea bass (Dicentrarchus labrax). Genet Sel Evol 47:68 Pearse DE, Janzen FJ, Avise JC (2002) Multiple paternity, sperm storage, and reproductive success of female and male painted turtles (Chrysemys picta) in nature. Behav Ecol Sociobiol 51:164–171 Pen I, Uller T, Feldmeyer B, Harts A, While GM, Wapstra E (2010) Climate-driven population divergence in sex-determining systems. Nature 468:436–438 Pérez P, de Los Campos G (2014) Genome-wide regression and prediction with the BGLR statistical package. Genetics 198:483–495 Rage JC (1998) Latest Cretaceous extinctions and environmental sex determination in reptiles. Bull Soc Géol Fr 169:479–483 Refsnider JM, Milne-Zelman C, Warner DA, Janzen FJ (2014) Population sex ratios under differing local climates in a reptile with environmental sex determination. Evol Ecol 28:977–989 Refsnider JM, Janzen FJ (2016) Temperature-dependent sex determination under rapid anthropogenic environmental change: evolution at a turtle’s pace? J Hered 107:61–70 Rhen T, Lang JW (1995) Phenotypic plasticity for growth in the common snapping turtle: effects of incubation temperature, clutch, and their interaction. Am Nat 146:726–747 Rhen T, Lang JW (1998) Among-family variation for environmental sex determination in reptiles. Evolution 52:1514–1520 Rhen T, Metzger K, Schroeder A, Woodward R (2007) Expression of putative sex-determining genes during the thermosensitive period of gonad development in the snapping turtle, Chelydra serpentina. Sex Dev 1:255–270 Rhodin AGJ, Stanford CB, Van Dijk PP, Eisemberg C, Luiselli L, Mittermeier RA et al. (2018) Global conservation status of turtles and tortoises (order Testudines). Chelonian Conserv Biol 17:135–161 Rao CR (1948) Large sample tests of statistical hypotheses concerning several parameters with applications to problems of estimation. Math Proc Cambridge Philos Soc 44:50–57 Ropp AJ, Reece KS, Snyder RA, Song J, Biesack EE, McDowell JR (2023) Fine-scale population structure of the northern hard clam (Mercenaria mercenaria) revealed by genome-wide SNP markers. Evol Appl 16:1422–1437 Roush D, Rhen T (2018) Developmental plasticity in reptiles: critical evaluation of the evidence for genetic and maternal effects on temperature-dependent sex determination. J Exp Zool A Ecol Integr Physiol 329:287–297 Sabath N, Itescu Y, Feldman A, Meiri S, Mayrose I, Valenzuela N (2016) Sex determination, longevity, and the birth and death of reptilian species. Ecol Evol 6:5207–5220 Saito T, Rehmsmeier M (2015) The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS One 10:e0118432 Samson J (2003) The life history strategy of a northern population of midland painted turtle, Chrysemys picta marginata. Dissertation, University of Guelph Santidrián Tomillo P (2022) When population-advantageous primary sex ratios are female-biased: changing concepts to facilitate climate change management in sea turtles. Clim Change 175:15 Santure AW, Garant D (2018) Wild GWAS—association mapping in natural populations. Mol Ecol Resour 18:729–738 Sarre SD, Georges A, Quinn A (2004) The ends of a continuum: genetic and temperature-dependent sex determination in reptiles. BioEssays 26:639–645 Silber S, Geisler JH, Bolortsetseg M (2011) Unexpected resilience of species with temperature-dependent sex determination at the Cretaceous–Palaeogene boundary. Biol Lett 7:295–298 Schroeder AL, Metzger KJ, Miller A, Rhen T (2016) A novel candidate gene for temperature-dependent sex determination in the common snapping turtle. Genetics 203:557–571. https://doi.org/10.1534/genetics.115.182840 Schwanz LE, Cordero GA, Charnov EL, Janzen FJ (2016) Sex-specific survival to maturity and the evolution of environmental sex determination. Evolution 70:329–341 Schwanz LE, Janzen FJ, Proulx SR (2010) Sex allocation based on relative and absolute condition. Evolution 64:1331–1345 Schwanz LE, Georges A, Holleley CE, Sarre SD (2020) Climate change, sex reversal and lability of sex-determining systems. J Evol Biol 33:270–281 Schwanz LE, Georges A (2021) Sexual development and the environment: conclusions from 40 years of theory. Sex Dev 15:7–22 Schwarzkopf L, Brooks RJ (1985) Sex determination in northern painted turtles: effect of incubation at constant and fluctuating temperatures. Can J Zool 63:2543–2547 Schwarzkopf L, Brooks RJ (1987) Nest-site selection and offspring sex ratio in painted turtles, Chrysemys picta. Copeia 1987:53–61 Shaffer HB, Minx P, Warren DE, Shedlock AM, Thomson RC, Valenzuela N et al. (2013) The western painted turtle genome, a model for the evolution of extreme physiological adaptations in a slowly evolving lineage. Genome Biol 14:R28 Souza CSD, Santos VSD, Martins Filho S (2024) Genomic prediction using the lmekin function from the coxme R package. Acta Sci Agron 46:e64243 Srivastava AK, Williams SM, Zhang G (2023) Heritability estimation approaches utilizing genome-wide data. Curr Protoc 3:e734 Staines MN, Versace H, Laloë JO, Smith CE, Madden Hof CA, Booth DT et al. (2023) Short-term resilience to climate-induced temperature increases for equatorial sea turtle populations. Glob Change Biol 29:6546–6557 Telemeco RS, Gangloff EJ, Cordero GA, Mitchell TS, Bodensteiner BL, Holden KG et al. (2016) Reptile embryos lack the opportunity to thermoregulate by moving within the egg. Am Nat 188:E13–E27 Thépot D (2021) Sex chromosomes and master sex-determining genes in turtles and other reptiles. Genes 12:1822 Thorsrud JA, Evans KM, Quigley KC, Srikanth K, Huson HJ (2025) Performance comparison of genomic best linear unbiased prediction and four machine learning models for estimating genomic breeding values in working dogs. Animals 15:408 Topping NE, Valenzuela N (2021) Turtle nest-site choice, anthropogenic challenges, and evolutionary potential for adaptation. Front Ecol Evol 9:808621 Turner SD (2014) qqman: an R package for visualizing GWAS results using QQ and manhattan plots. bioRxiv 005165 Uffelmann E, Huang QQ, Munung NS, De Vries J, Okada Y, Martin AR et al. (2021) Genome-wide association studies. Nat Rev Methods Primers 1:59 VanRaden PM (2008) Efficient methods to compute genomic predictions. J Dairy Sci 91:4414–4423 Velo-Antón G, Becker CG, Cordero-Rivera A (2011) Turtle carapace anomalies: the roles of genetic diversity and environment. PLoS One 6:e18714 Wang J, Zhou Z, Zhang Z, Li H, Liu D, Zhang Q et al. (2018) Expanding the BLUP alphabet for genomic prediction adaptable to the genetic architectures of complex traits. Heredity 121:648–662 Warner DA, Shine R (2008) The adaptive significance of temperature-dependent sex determination in a reptile. Nature 451:566–568 Wickham H (2011) ggplot2. Wiley Interdiscip Rev Comput Stat 3:180–185 Wnek JP, Bien WF, Avery HW (2013) Artificial nesting habitats as a conservation strategy for turtle populations experiencing global change. Integr Zool 8:209–221 Wray NR, Yang J, Goddard ME, Visscher PM (2010) The genetic interpretation of area under the ROC curve in genomic profiling. PLoS Genet 6:e1000864 Yang J, Weedon MN, Purcell S, Lettre G, Estrada K, Willer CJ et al. (2011) Genomic inflation factors under polygenic inheritance. Eur J Hum Genet 19:807–812 Yarberry W (2021) Dplyr. In: CRAN recipes: DPLYR, stringr, lubridate, and regex in R. Apress, Berkeley, CA, pp 1–58 Yntema CL (1979) Temperature levels and periods of sex determination during incubation of eggs of Chelydra serpentina. J Morphol 159:17–27 Yntema CL, Mrosovsky N (1980) Sexual differentiation in hatchling loggerheads (Caretta caretta) incubated at different controlled temperatures. Herpetologica 36:33–36 Additional Declarations There is no duality of interest Cite Share Download PDF Status: Under Review Version 1 posted Review # 1 received at journal 12 May, 2026 Reviewer # 2 agreed at journal 22 Apr, 2026 Reviewer # 1 agreed at journal 22 Apr, 2026 Reviewers invited by journal 22 Apr, 2026 Editor assigned by journal 05 Apr, 2026 First submitted to journal 05 Apr, 2026 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-9324153","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":617702855,"identity":"d26bad2d-8185-43b3-8716-7dd4ad3ba72a","order_by":0,"name":"Robin Lloyd","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABC0lEQVRIiWNgGAWjYFACHoYDDAwWMkCW4QMGBgkZYrVI8ABZxgYQBjNYXAKfFgaoFjMJCJeAFvP23oOHC4Ba+CWSt1XdqLHgYZDIP/i4ouJOHT8D88MPWLTInDmXcHgGUIvkjLSy2znHgNZJJDMbnjnzTEKygc0Ym1USEjkGh4HKeAxu5JjdzmEDa2GTbGw7LGFwgAer61C0FOf8Q9XC/IOQFubcNlQtbFht4QH6hccA6JeeZ8XSuX0SPGw8j40NG84clpzZzGZmgU0Le+/hzzwVNnL87MkbP+d8qwMyEh8+bKg4zM/P3vz4Bq6QZgDGIYNAAoTNBhdlxqkeCvgPEFIxCkbBKBgFIxUAADE+UXTmIKxaAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-8487-2102","institution":"University of Toronto","correspondingAuthor":true,"prefix":"","firstName":"Robin","middleName":"","lastName":"Lloyd","suffix":""},{"id":617702856,"identity":"e465e410-01ad-4102-b07b-897293b1fd56","order_by":1,"name":"Melanie Massey","email":"","orcid":"","institution":"University of Toronto","correspondingAuthor":false,"prefix":"","firstName":"Melanie","middleName":"","lastName":"Massey","suffix":""},{"id":617702857,"identity":"617ed507-b34c-407b-96be-f23100290746","order_by":2,"name":"Joanna Rifkin","email":"","orcid":"https://orcid.org/0000-0003-1980-5557","institution":"University of Toronto","correspondingAuthor":false,"prefix":"","firstName":"Joanna","middleName":"","lastName":"Rifkin","suffix":""},{"id":617702858,"identity":"5c79f83f-2994-4956-aaa1-e192981374fe","order_by":3,"name":"Jacqueline Litzgus","email":"","orcid":"","institution":"Laurentian University","correspondingAuthor":false,"prefix":"","firstName":"Jacqueline","middleName":"","lastName":"Litzgus","suffix":""},{"id":617702859,"identity":"ed889a55-6bac-47e6-9063-67e67036bcbb","order_by":4,"name":"Njal Rollinson","email":"","orcid":"","institution":"University of Toronto","correspondingAuthor":false,"prefix":"","firstName":"Njal","middleName":"","lastName":"Rollinson","suffix":""}],"badges":[],"createdAt":"2026-04-05 05:10:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9324153/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9324153/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108206984,"identity":"3ad4b644-988e-4eda-a5eb-0c79a4262ea6","added_by":"auto","created_at":"2026-04-30 13:04:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2664902,"visible":true,"origin":"","legend":"\u003cp\u003eLong-term monitoring field site ponds and subject model organism within Algonquin Provincial Park, Ontario, Canada. (A) Wolf Howl Pond bog-mat habitat, where turtles were frequently captured while basking in May 2022 during initial mark–recapture surveys. (B) West Rose Lake spruce bog viewed from the nesting embankment in June 2022 prior to egg collection. (C) Typical adult female midland painted turtle (\u003cem\u003eChrysemys picta marginata\u003c/em\u003e) post-nesting at West Rose Lake in June 2022. Photos taken by Robin Lloyd.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-9324153/v1/73fdecbc2c39a8441c3fc7f5.png"},{"id":108206985,"identity":"e25ac1b2-4ad1-4de7-82e6-2edb82c51190","added_by":"auto","created_at":"2026-04-30 13:04:56","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":806041,"visible":true,"origin":"","legend":"\u003cp\u003eGenome-wide association study (GWAS) results for sex in wild adult painted turtles. Manhattan plots show the –log₁₀(\u003cem\u003eP\u003c/em\u003e) values for SNPs across 25 chromosomes, with no loci reaching genome-wide significance (–log₁₀(P) \u0026gt; 8 threshold).\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-9324153/v1/ba70ac099ced947e7c75793f.png"},{"id":108206986,"identity":"25f77f9a-e107-4731-ad16-16c2b0781c8b","added_by":"auto","created_at":"2026-04-30 13:04:57","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1183066,"visible":true,"origin":"","legend":"\u003cp\u003eGenome-wide association study (GWAS) results for sex in experimental hatchling painted turtles. Manhattan plots show the –log₁₀(\u003cem\u003eP\u003c/em\u003e) values for SNPs across 25 chromosomes, with no loci reaching genome-wide significance (–log₁₀(P) \u0026gt; 8 threshold).\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-9324153/v1/d76095b2cc75e33622984e16.png"},{"id":108803897,"identity":"c3f90418-8455-47cc-804d-01227d1ef3e3","added_by":"auto","created_at":"2026-05-08 15:10:15","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":353136,"visible":true,"origin":"","legend":"\u003cp\u003eQuantile–quantile (QQ) plots from the wild adult (A) and hatchling (B) painted turtle genome-wide association study (GWAS) comparing observed and expected –log₁₀(\u003cem\u003eP\u003c/em\u003e) values with the shaded area representing 95% confidence intervals. Overall, the QQ-plots suggest appropriate calibration of GWAS test statistics, with genomic inflation factors (λ) close to 1 in both datasets.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-9324153/v1/047777e7544eb75c7d1b171d.png"},{"id":108206987,"identity":"eef144db-0a92-4240-a49e-b26284abf0fd","added_by":"auto","created_at":"2026-04-30 13:04:57","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":238245,"visible":true,"origin":"","legend":"\u003cp\u003eGenomic prediction results for sex classification in adult and hatchling painted turtles using genomic best linear unbiased prediction (GBLUP) models with 10,423 SNPs and 500 bootstrap replicates. Panels show (A) prediction accuracy, (B) area under the curve (AUC), (C) F1 score, and (D) narrow-sense heritability (\u003cem\u003eh²\u003c/em\u003e). Points mark the mean values, and error bars show the 95% confidence intervals across replicates.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-9324153/v1/4c5bcf3d914600cb3f653b2e.png"},{"id":108809089,"identity":"ad9590e6-0260-4a15-80aa-f01213809739","added_by":"auto","created_at":"2026-05-08 15:49:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5581035,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9324153/v1/0738a7f9-b592-4132-99c3-afd1e9bb4e38.pdf"}],"financialInterests":"There is no duality of interest","formattedTitle":"Genetic influence on the sex ratio of a turtle with temperature-dependent sex determination","fulltext":[{"header":"Introduction ","content":"\u003cp\u003eSex ratio has long been recognized as a central driver of population dynamics (Fisher, 1930; Bull, 1983; Bull and Charnov, 1989). Theory predicts that natural populations should exhibit a balanced adult sex ratio, as the genetic contribution of each sex to the next generation must be equal. Deviation from a balanced sex ratio would favour overproduction of the rarer sex, and frequency-dependent selection would ultimately stabilize the sex ratio over time (Fisher, 1930; Charnov and Bull, 1977; Bull and Charnov, 1989). Most organisms exhibit genotypic sex determination (GSD), where sex is determined genetically, often through sex chromosomes or sex-determining genes (Bachtrog et al., 2014; Vandeputte, 2016). Species with GSD typically maintain a balanced sex ratio, or nearly so, from birth to adulthood (Bachtrog et al., 2014;\u0026nbsp;B\u0026oacute;kony et al., 2019). However, many organisms have environmental sex determination (ESD), where sex depends on external environmental factors (\u003cem\u003ee.g\u003c/em\u003e., temperature, social cues, salinity) that act after fertilization, typically during embryonic development (Bachtrog et al., 2014; B\u0026oacute;kony et al., 2019). Species with ESD frequently deviate from a 1:1 sex ratio at birth and in adulthood, which challenges Fisher\u0026rsquo;s (1930) theoretical predictions of balanced sex ratios (Bull and Charnov, 1988; Charnov and Bull, 1989). Over the last several decades, unbalanced sex ratios of ESD species have prompted extensive investigation into the adaptive significance of various forms of ESD, accompanied by growing concerns over the potential impacts of rapid environmental change on population sex ratios (Charnov and Bull, 1977; Janzen and Phillips, 2006; Warner and Shine, 2008; Pen et al., 2010).\u003c/p\u003e\n\u003cp\u003eThe most widespread form of ESD is temperature-dependent sex determination (TSD), which is commonly found in reptiles, as well as some fishes and invertebrates (Bull, 1980; Ewert et al., 2004). Under TSD, incubation temperature influences gonadal differentiation into ovaries or testes during the thermosensitive period (TSP), which, in reptiles, typically occurs around the middle third of embryonic development (Yntema, 1979; Bull, 1980; Bull, 1983; Janzen and Paukstis, 1991). Under TSD, constant incubation at an extreme temperature may lead to production of a single sex; for instance, in many TSD turtles, males are produced at cool temperatures and females at warm temperatures (see\u0026nbsp;Lockley and Eizaguirre, 2021). At intermediate temperatures there is a transitional range of temperatures (TRT) where both sexes are produced and sex ratios are mixed. The TRT represents a temperature range where there is a transition between an overproduction of one sex to an overproduction of the other sex, with a pivotal temperature (Tpiv) that produces a 1:1 sex ratio (Mrosovsky and Yntema, 1980; Mrosovsky and Pieau, 1991; Lockley and Eizaguirre, 2021). Variation in Tpiv is influenced by maternal effects (Ewert et al., 2004; Schwanz et al., 2010; Carter et al., 2016; Roush and Rhen, 2018) but also has a genetic basis (Janzen and Paukstis, 1991; Mrosovsky and Pieau, 1991; Janzen, 1992; Ewert et al., 2004), and appears in some cases to have evolved across latitudinal gradients, suggesting adaptation to local conditions (Bull et al., 1982; Ewert et al., 2005; Li et al., 2022).\u003c/p\u003e\n\u003cp\u003eIn GSD systems, the bipotential gonad develops into either testes or ovaries depending on the presence of a master sex determining gene such as \u003cem\u003eDmrt1\u003c/em\u003e in birds or \u003cem\u003eSry\u003c/em\u003e in mammals (Tezak et al., 2020; Th\u0026eacute;pot, 2021). There is no clear evidence of a master sex determining gene in TSD reptiles, but gene regulation underlies temperature-driven gonadal fate: \u003cem\u003eDmrt1\u003c/em\u003e and \u003cem\u003eSox9\u003c/em\u003e mRNA expression increase at male-producing temperatures, whereas aromatase, androgen receptor (\u003cem\u003eAr\u003c/em\u003e), and \u003cem\u003eFoxl2\u0026nbsp;\u003c/em\u003eare upregulated at female-producing temperatures (Schroeder et al., 2016; Matsumoto and Crews, 2017; Roush and Rhen, 2018; Th\u0026eacute;pot, 2021). Given that Tpiv exhibits heritability in TSD species (Janzen, 1992; McGaugh and Janzen, 2011; Refsnider and Janzen, 2016), it is likely that genes regulating hormone expression act as proximate mechanisms through which genetic variation in physiological traits contributes to putatively adaptive responses of Tpiv across geographic gradients. Yet, direct genotype\u0026ndash;sex associations remain understudied in TSD reptiles (Roush and Rhen, 2018). Most efforts to link genotype with sex have relied on candidate gene approaches, which have produced mixed results. For example, Schroeder et al. (2016) identified an association between the cold-induced RNA binding protein (\u003cem\u003eCIRBP\u003c/em\u003e) gene and sex in hatchling common snapping turtles (\u003cem\u003eChelydra serpentina\u003c/em\u003e), suggesting that genotype may influence sensitivity to temperature during sex determination. By contrast, Matsumoto and Crews (2017) found no association between polymorphisms in the aromatase (\u003cem\u003ecyp19a1\u003c/em\u003e) gene and sex in pond sliders (\u003cem\u003eTrachemys scripta\u003c/em\u003e), indicating that some key regulators are highly conserved and primarily regulated by temperature rather than genetic variation. A key gap in understanding the genetic basis of TSD is the lack of genome-wide analyses testing for sex-specific associations under controlled incubation conditions, assessing the adaptive potential of physiological thresholds such as Tpiv (Roush and Rhen, 2018; Th\u0026eacute;pot, 2021), and quantifying heritability of sex (Janzen, 1992; McGaugh and Janzen, 2011; Refsnider and Janzen, 2016).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMany TSD species may be vulnerable to the threat of rapid climate change, where a sustained warming will bias sex ratios, potentially resulting in population crashes over the next century (Janzen, 1994b; Schwanz et al., 2020; Tomillo, 2022; Staines et al., 2023). Yet, TSD has persisted for over 200 million years, surviving major climatic upheavals, suggesting TSD is not necessarily maladaptive under climate change (Rage, 1998; McGaugh and Janzen, 2011; Silber et al., 2011). Even though sex ratios in TSD species frequently deviate from a 1:1 sex ratio, Fisherian sex-ratio selection (Fisher, 1930) may become increasingly important under rapid climate shifts for stabilizing sex ratios, with genetic variation playing a critical role in responses to stabilizing selection (McGaugh and Janzen, 2011). Adaptive evolutionary responses of traits associated with sex ratio may arise in heritable thermal physiological traits (\u003cem\u003ee.g.\u003c/em\u003e, Tpiv, TRT, TSP) (Mitchell and Janzen, 2010; McGaugh and Janzen, 2011; Krueger et al., 2025), and/or heritable nest-site choice behaviour (McGaugh et al., 2010; Refsnider and Janzen, 2016). Of these, heritable physiological traits are predicted to be the most effective avenue for TSD species to adapt to rapid climate change (McGaugh and Janzen, 2011; Crowther and Schwanz, 2025). This is because developing embryos have little to no capacity to behaviourally thermoregulate (Telemeco et al., 2016), such that natural selection is expected to act directly on variation in their physiology. Heritable nest-site choice, on the other hand, is constrained by the availability of suitable nesting habitat (McGaugh and Janzen, 2011), making changes in nest-site choice much slower than the pace of climate change, and ultimately of potentially limited adaptive potential except under climate extremes (McGaugh and Janzen, 2011; Topping and Valenzuela, 2021; Crowther and Schwanz, 2025).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhile genetic variation for physiological traits, especially traits related to the Tpiv and TRT, may hold the adaptive potential for TSD species to persist under climate change, the scope of these genetic influences to mediate sex-ratio bias under climate change is unclear (McGaugh et al., 2010; Roush and Rhen, 2018; Crowther and Schwanz, 2025). Heritability of Tpiv and TRT exists in TSD species (reviewed in Refsnider and Janzen, 2016), but heritable variation is not expressed at extreme incubation temperatures, where the influence of temperature overrides any genetic influence on sex ratios that is otherwise found at intermediate temperatures (Roush and Rhen, 2018; Crowther and Schwanz, 2025). Additive genetic variation for Tpiv under extreme climatic conditions is therefore negligible, and a rapid increase in incubation temperatures under climate change may outpace the potential for an adaptive response of TSD species. A key step in understanding the threat posed by climate change is understanding the extent of genetic variation for sex in wild, unmanipulated TSD populations, especially those with sex ratios that have already undergone a thermally induced bias. Under sustained exposure to extreme incubation temperatures, wild populations with biased sex ratios might be expected to exhibit less genetic variation. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWild adults, shaped by long-term natural selection, provide a snapshot of whether genetic effects on sex persist in the environment, and may offer insight into the adaptive potential of TSD traits under climate change. Yet, in the wild, genetic signals can be confounded by environmental heterogeneity and maternal influences such as yolk steroids, which can also affect sex ratios (Bowden et al., 2000; Roush and Rhen, 2018). In contrast, embryos incubated at the Tpiv undergo sex determination \u0026nbsp; when environmental variance is minimized, allowing genetic effects on sex to be more directly expressed. Here, we compare wild adult turtles with TSD to hatchlings incubated at Tpiv, estimating heritability of sex in the wild and in the lab, and testing whether genetic associations with sex are robust in natural populations or masked by environmental noise\u003cstrong\u003e.\u0026nbsp;\u003c/strong\u003eThe main objectives of this study are to determine whether adult-wild and Tpiv-reared turtles with TSD show genomic regions associated with sex, whether sex has a detectable additive genetic basis, and how accurately sex can be predicted from genetic variation. Our comparative framework provides a critical step toward evaluating the adaptive potential of TSD species under rapid climate change, investigating specific loci associated with sex, and detection of genetic signal in a complex trait.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eStudy Site and Population\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e \u0026ndash; The present study is part of a long-term study on midland painted turtle (\u003cem\u003eChrysemys picta marginata\u003c/em\u003e) initiated in 1978 at Wolf Howl Pond (WHP) and West Rose Lake (WRL), two black spruce (\u003cem\u003ePicea mariana\u003c/em\u003e) bogs in Algonquin Provincial Park, Ontario, Canada (WHP: 45\u0026deg;34\u0026prime;00\u0026Prime;N, 78\u0026deg;41\u0026prime;00\u0026Prime;W) (see Figure 1). Both lakes are bisected by a 230 m long, 6\u0026ndash;7 m wide decommissioned railway embankment built in the 1890s, now part of a park hiking trail. The embankment provides a warm, sparsely vegetated nesting habitat that has been monitored annually for turtle nesting activity since 1990 (Hughes and Brooks, 2006). Painted turtles are a medium-sized freshwater turtle exhibiting male-female pattern TSD, where cooler incubation temperatures produce males and warmer temperatures produce females (see Lockley and Eizaguirre, 2021).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTurtles are captured annually by canoe and dipnet in spring (May) and by hand during the nesting season (June \u0026ndash; early July). Individuals are permanently marked with unique notches filed into marginal scutes (Cagle, 1939), and in some cases fitted with Passive Integrated Transponder (PIT) tags. Sex is determined using external morphology: males have elongated foreclaws, a cloacal opening farther from the base of the tail, and smaller body size, while females are larger and lack these traits (Ernst and Lovich, 2009; Moldowan et al., 2016, 2017). Males typically reach sexual maturity at 7-10 years (~90 mm carapace length), and females at 10\u0026ndash;16 years (~110 mm) within this population (Samson, 2003; Schwarzkopf and Brooks, 1987) with a lifespan exceeding 60 years for many individuals. At WHP, turtle density is approximately 129 individuals per hectare (Rouleau, 2020), with a strongly female-biased sex ratio of 0.29 males to 1 female (Samson, 2003; Rouleau, 2020; Moldowan et al., 2020). Males frequently sire offspring across multiple clutches, and females are capable of storing sperm for several years (Pearse et al., 2002; Hughes, 2011). Blood sampling of adults for genomic sequencing was conducted from 2018 \u0026ndash; 2022, with samples collected from the subcarapacial vein and preserved on Whatman\u0026reg; FTA\u0026reg; cards.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003e2022 Fieldwork and Incubation Methods\u003c/em\u003e\u003c/strong\u003e \u0026ndash; We collected eggs from 133 painted turtle clutches between 1 and 25 June 2022. Once a female had nested, she was identified, and three eggs were randomly selected per clutch (mean clutch size \u0026asymp; 6.7 eggs) using a randomizer smartphone app (Random: All Things Generator, version 2.10.16). All eggs were weighed and measured at the Algonquin Wildlife Research Station (AWRS). The three randomly selected eggs were transferred to 20 cm x 10 cm plastic containers, half-filled with moist vermiculite (equal parts water and vermiculite), with puncture holes in the lid for airflow. Containers were placed in one of two IN30 Torrey Pines Echo Therm incubators at 27.3\u0026deg;C (the Tpiv for this population, Schwarzkopf and Brooks, 1985, 1987), for approximately three weeks, until the end of the nesting season. Eggs were moved weekly within and between the two incubators to disrupt thermal gradients. To minimize water loss, we compared the current weight of each container to its initial weight every 3-4 days, replenishing evaporated water to each container. After the nesting season, all eggs were transferred to the University of Toronto in an automobile over three hours, and upon arrival at the University, eggs were immediately placed into two Memmert Natural Convection incubators, under the same protocol as used at the AWRS until hatching. Upon hatching, each hatchling was humanely euthanized using an injection of 10 mg/ml Alfaxalone, and 1-2 minutes after the injection, hatchlings were decapitated and immediately pithed. A blood sample was taken from each hatchling for genetic analysis using Whatman\u0026reg; FTA\u0026reg; cards. The sex of the hatchlings was determined via dissection under an Olympus SZX16 dissection microscope. Sex determination was conducted three times by two researchers. The first determination was performed by R.B.L and occurred during the dissection of each hatchling. The second and third determinations were made by R.B.L. and M.D.M. using photos of the Dissected hatchlings. Sex determination was based on the morphology of the gonads: male gonads were small, short, and rounded, while female gonads were long and thin (Yntema and Mrosovsky, 1980; Wyneken et al., 2007).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eFiltering of SNP Data\u003c/em\u003e\u0026nbsp;\u003c/strong\u003e\u0026ndash; Blood samples collected from wild adult turtles and experimentally incubated hatchlings were genotyped for single nucleotide polymorphisms (SNPs) at Diversity Arrays Technology (DarTSeq, Canberra, Australia), which targets hypomethylated, low-copy genomic regions via complexity reduction methodology (see Melville et al., 2017). The DarTSeq platform used SNP loci that were called and aligned to a western painted turtle (\u003cem\u003eChrysemys picta belli\u003c/em\u003e) reference genome (GCF_00241765.4) to ensure consistent marker positioning across individuals (Shaffer et al., 2013). We filtered raw SNP data (82,522 loci) from DarTSeq to improve data quality and identify informative SNPs using the R package \u003cem\u003edartR\u003c/em\u003e (v4.3.2; Gruber et al., 2018; R Core Team, 2018). Following the SNP filtering framework described by Gruber et al. (2019), we first removed monomorphic loci, which lack allelic variation and thus provide no information for genetic analyses. We next retained only SNPs with a repeatability \u0026ge; 0.95 to ensure consistent detection across replicates and improve genotyping reliability. \u0026nbsp;We further excluded secondary loci\u0026mdash;multiple SNPs from the same sequence fragment\u0026mdash;to eliminate redundant information. We then applied a call rate threshold of 0.95, removing loci with more than 5% missing data (Anderson et al., 2010). SNPs with minor allele frequency \u0026lt; 0.01 were removed, as rare alleles offer limited statistical power and may introduce noise (reference). We also removed loci that significantly deviated from Hardy\u0026ndash;Weinberg equilibrium (𝛼 \u0026lt; 0.0001), where the p-value reflects the likelihood that deviations in genotype frequencies arise by chance. This stringent threshold is commonly used to eliminate SNPs potentially affected by genotyping errors, thereby improving data quality for downstream analyses (Chen et al., 2017; Ropp et al., 2023). Finally, we filtered out individual turtles with a call rate below 0.95, meaning those missing more than 5% of SNP genotype calls, to ensure that only samples with high genotyping completeness were retained for analysis (Anderson et al., 2010). Approximately 0.63% of SNP genotype calls were missing in the filtered wild adult turtle dataset and 0.65% in the filtered experimental hatchling dataset. To replace missing SNP genotype calls, the most common allele (mode) for each SNP was imputed using the \u003cem\u003eDplyr\u003c/em\u003e package in R, as the genomic analyses required complete genotype data with no missing values (Yarberry, 2021; Appadurai et al., 2023).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eGenome-wide Association Study (GWAS)\u0026nbsp;\u003c/em\u003e\u0026ndash;\u0026nbsp;\u003c/strong\u003eWe conducted two separate genome-wide association studies (GWAS) to identify SNPs associated with sex: one for wild adult turtles and one for experimental hatchlings. We separated these groups because adults incubated naturally and were likely subject to post-hatching selection, while hatchlings were reared under controlled, constant conditions at the Tpiv. We coded the filtered SNP data as 0, 1, or 2, representing the number of alternate alleles per locus. The phenotypes were binary-coded as 1 (male) and 0 (female). To control for relatedness and population structure, we used the \u003cem\u003eGMMAT\u003c/em\u003e package in R (Chen et al., 2019) to fit a logistic mixed null model of the form: \u003cem\u003elogit(P(sex = 1)) = \u0026beta;₀ + u\u003c/em\u003e, where \u003cem\u003e\u0026beta;₀\u003c/em\u003e is the fixed intercept and \u003cem\u003eu\u003c/em\u003e is a random effect accounting for kinship among individuals. Relatedness was modeled using a genomic relationship matrix (GRM) constructed following VanRaden \u0026nbsp;(2008) where genotype values were centered by allele frequency and scaled by the total expected heterozygosity across loci to quantify pairwise genomic similarity among individuals. This GRM or kinship matrix was used to model the random genetic effect in the mixed model. The logistic mixed null model was first fit without SNP effects, and each SNP was then tested individually using a score test via the \u003cem\u003elm.score\u003c/em\u003e function in R to assess its association with sex while accounting for relatedness (Rao, 1948; Chen et al., 2016). We then merged \u003cem\u003eP\u003c/em\u003e-value results with SNP metadata (chromosome number and genomic position) provided by Diversity Arrays Technology for genomic mapping. For visualization, we then used the \u003cem\u003eqqman\u003c/em\u003e (Turner, 2014) and \u003cem\u003eggplot2\u003c/em\u003e packages (Wickham, 2011) to generate Manhattan and QQ plots to showcase GWAS results. To evaluate potential test statistic inflation, we calculated the genomic inflation factor (\u003cem\u003e\u0026lambda;\u003c/em\u003e), defined as: \u003cem\u003e\u0026lambda; = median(\u0026chi;\u0026sup2;_observed) / 0.456,\u003c/em\u003e where 0.456 is the expected median of a \u003cem\u003e\u0026chi;\u0026sup2;\u003c/em\u003e distribution with 1 degree of freedom under the null hypothesis (Yang et al., 2011). The \u003cem\u003e\u0026lambda;\u003c/em\u003e value was embedded in each QQ plot to assess the adequacy of population structure correction.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eGenomic Prediction via Genomic Best Linear Unbiased Prediction (GBLUP)\u003c/em\u003e \u0026ndash; Genomic prediction was used to assess the extent to which genome-wide SNP variation explains sex in our painted turtle population. We implemented genomic prediction using the \u003cem\u003eBGLR\u003c/em\u003e package in R (P\u0026eacute;rez and Campos, 2014), applying a Reproducing Kernel Hilbert Space (RKHS) model with a linear genomic kernel, which is equivalent to a standard GBLUP model that uses a GRM to account for additive genetic effects (VanRaden, 2008; Gianola and Kaam, 2008; Wang et al., 2018). The GRM was constructed using the same VanRaden (2008) approach as in the GWAS analyses and used as the genomic kernel in the GBLUP models to account for genome-wide relatedness among individuals. Our GBLUP model was defined as \u003cem\u003ey = \u0026mu; + u + \u0026epsilon;\u003c/em\u003e, where \u003cem\u003ey\u003c/em\u003e is the vector of binary sex phenotypes (1 = male, 0 = female), \u003cem\u003e\u0026mu;\u003c/em\u003e is the overall mean, \u003cem\u003eu\u003c/em\u003e is the vector of additive genetic effects assumed to follow a normal distribution \u003cem\u003eu ~ N(0, G\u0026sigma;\u0026sup2;ᵤ)\u003c/em\u003e, and \u003cem\u003e\u0026epsilon;\u003c/em\u003e is the residual error, assumed to follow \u003cem\u003e\u0026epsilon; ~ N(0, I\u0026sigma;\u0026sup2;ₑ)\u0026nbsp;\u003c/em\u003e(VanRaden, 2008; Wang et al., 2018; Souza et al., 2024). Bayesian estimation of model parameters was conducted using Gibbs sampling with 3,000 iterations and a 1,000-iteration burn-in (P\u0026eacute;rez and Campos, 2014). To evaluate model performance while accounting for sex ratio imbalance, we conducted 500 bootstrap iterations separately for adults and hatchlings, where sampling was stratified to preserve the original sex ratio of the sample. In each iteration, 85% of individuals were randomly selected to form the training set, which was used to fit the prediction model and construct the genomic kernel. The remaining 15% of individuals comprised the test set, which was completely held out during model training. Sex for test individuals was predicted based on their genomic similarity to the training individuals.\u003c/p\u003e\n\u003cp\u003eModel performance was evaluated using four metrics: (1) accuracy, defined as the proportion of correct classifications across males and females (Dekkers and Cheng, 2021); (2) the area under the receiver operating characteristic curve (AUC\u003cstrong\u003e)\u003c/strong\u003e, which measures how well the model distinguishes between males and females across all possible probability thresholds used for classification; values near 0.5 indicate random prediction, while values near 1.0 indicate perfect discrimination (Wray et al., 2010); (3) F1 score, the harmonic mean of precision (the proportion of predicted males that are truly male) divided by recall (the proportion of actual males correctly predicted), which balances false positives and false negatives and is especially informative when sex classes are imbalanced (Saito and Rehmsmeier, 2015); and (4) narrow-sense heritability (\u003cem\u003eh\u0026sup2;\u003c/em\u003e), which quantifies the proportion of phenotypic variance attributable to additive genetic effects (Falconer, 1996; Holland et al., 2003; Srivastava et al., 2023). We estimated narrow-sense heritability on the observed phenotypic scale for each GBLUP model iteration within each bootstrap replicate by calculating \u003cem\u003eh\u0026sup2; =\u003c/em\u003e \u003cem\u003eVar(u) / [Var(u) + Var(\u0026epsilon;)]\u003c/em\u003e, where \u003cem\u003eVar(u)\u003c/em\u003e is the variance of genomic breeding values and \u003cem\u003eVar(\u0026epsilon;)\u003c/em\u003e is the residual variance estimated by the GBLUP model (Holland et al., 2003; VanRaden, 2008; Souza et al., 2024). To quantify uncertainty, we calculated 95% confidence intervals for each performance metric and heritability estimate using a nonparametric percentile bootstrap approach based on the empirical distribution across the 500 bootstrap replicates.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eA total of 399 eggs were collected from the study population. Of these, 327 (82%) hatched successfully after incubation at the University of Toronto. Eighteen hatchlings were excluded because sex could not be determined, either due to gonadal tearing during dissection or because the gonads were too underdeveloped to classify. Of the remaining 309 hatchlings, 13 were not confidently sexed from photographs, leaving 296 individuals with confirmed sex for genetic analysis. An additional 25 individuals were removed because of incomplete or failed genotyping, resulting in a final hatchling dataset of 271 turtles (111 females, 160 males) representing 125 distinct mothers. Separately, genotyping was also conducted on wild-caught adult turtles sampled in 2022 and from previous years. After SNP filtering and quality control, the final dataset included 839 individuals and 10,423 SNPs comprised of the 271 hatchlings and 568 wild-caught turtles. Among the wild-caught individuals, sex and genotyping data were successfully obtained for 474 adults (389 females, 85 males) including all mothers of the incubated hatchlings, which were used in downstream genomic analyses.\u003c/p\u003e\n\u003cp\u003eSeparate GWAS analyses were conducted for adult and hatchling painted turtles to identify SNPs associated with sex. In both datasets, SNPs were distributed across all 25 chromosomes, and no loci reached conventional genome-wide significance thresholds (Figures 2 and 3). QQ plots indicated minimal inflation in test statistics, with \u0026lambda; = 1.017 for adults and 1.021 for hatchlings (Figure 4a,b), suggesting appropriate control for population structure and relatedness.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo assess the predictive power of genome-wide SNPs for sex determination, GBLUP models were applied to both adult and hatchling datasets using 10,423 filtered SNPs across 500 bootstrap replicates. Prediction accuracy was higher in adults than in hatchlings. In adults, the mean prediction accuracy (95% CI) was 0.757 (0.637\u0026ndash;0.855), with an average AUC of 0.661 (0.468\u0026ndash;0.805) (Figure 5a,b). The F1 score was relatively low at 0.351 (0.135\u0026ndash;0.561), reflecting a trade-off between precision and recall, which was likely influenced by the imbalanced sex ratio in the adult sample (Figure 5c). The estimated narrow-sense heritability of sex in adults was 0.29 (0.203\u0026ndash;0.391), suggesting that roughly one-third of the variance in sex classification could be attributed to additive genetic effects in the wild population (Figure 5d).\u003c/p\u003e\n\u003cp\u003eThe hatchling dataset featured lower predictive performance. The mean prediction accuracy was 0.571 (0.426\u0026ndash;0.707), and the AUC was 0.568 (0.452\u0026ndash;0.704) (Figure 5a,b). Compared to adults, the F1 score was substantially (but not significantly) greater in hatchlings at 0.662 (0.521\u0026ndash;0.779), likely reflecting the more balanced sex ratio in this group compared to the wild adults and improved recall for the minority class (Figure 5c). The estimated heritability of sex in hatchlings was 0.16 (0.114\u0026ndash;0.237), slightly but not significantly less than that of the adult wild population (Figure 5d).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eAdaptive responses of TSD species to warmer incubation conditions will depend on the microevolution of traits underlying sex determination and the rate at which these traits can evolve relative to environmental change (McGaugh and Janzen, 2011; Refsnider and Janzen, 2016; Topping and Valenzuela, 2021; Krueger et al., 2025). Here we present, to our knowledge, the first within-population comparison of genotype\u0026ndash;sex associations between wild adults and hatchlings incubated at the Tpiv in a TSD species. In adults, we detected modest heritability of sex, whereas hatchlings exhibited a slightly (but not significantly) lower heritability, and no single SNP locus was a strong predictor of sex for hatchlings or adults. These results are consistent with a polygenic basis for sex determination (Roush and Rhen, 2018), where sex is influenced by the cumulative effects of many small-effect loci, with no evidence for a significant difference between wild adults and hatchlings incubated at the Tpiv.\u003c/p\u003e\n\u003cp\u003eIn the present study, controlled incubation of hatchlings at the Tpiv should have minimized environmental variance and enhanced genetic signal for sex, but we observed no such pattern (Figure 5). One explanation for why hatchlings did not exhibit greater heritability for sex than adults is that genotype-sex associations in wild adults are the outcome of both post-hatching survival and developmental conditions of the embryo. Post-hatching filtering through ontogeny reflects viability selection (Hadfield, 2008), where mortality after hatching disproportionately removes individuals whose genotype\u0026ndash;sex combinations are maladaptive, consequently increasing the observed genetic signal by virtue of better-aligning genotype with sex (McGaugh et al., 2011; Mittell et al., 2025). Sex-linked differences in embryonic thermal tolerance (Tomillo, 2022), hormone expression (Crews et al., 1989; Rhen et al., 2007), or metabolic requirements (O\u0026rsquo;Steen and Janzen, 1999) could drive such filtering. In addition, some of the wild adults may be siblings that developed in the same nesting environment, which can align similar genotypes with sex even when no genetic variation for sex was expressed, giving the appearance of stronger heritable effects when causal loci are weak (McGaugh et al., 2010). While the GRMs in our models account for genomic similarities, they do not explicitly model maternal effects and shared environmental conditions that could align genotype and incubation environment across generations, potentially leading to modest inflation of heritability estimates in wild adults. By contrast, incubation of embryos precisely at the Tpiv reduces between-clutch variation in incubation temperature (Rhen and Lang, 1995), standardizing developmental conditions around the transitional range of temperature, such that sexual outcomes would be most sensitive to maternal effects (e.g., yolk hormones) and genotype (Ewert et al., 2004; Roush and Rhen, 2018) than in wild incubated nests. The absence of viability selection (or weaker viability selection under modest embryo mortality), maternal nest-site choice, and site fidelity in this controlled laboratory setting may reduce genotype\u0026ndash;sex correlation, leading to uninflated heritability estimates. Finally, methodological constraints may have also dampened estimates of heritability for hatchlings, as hatchling turtles can be challenging to sex based on gonadal examination (Yntema and Mrosovsky, 1980), leading to mistaken sex classification that dampens estimates of heritability for sex.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe did not detect significant SNPs for sex in our GWAS, which is consistent with findings in other TSD systems. While our sample sizes were among the largest yet reported for a reptile GWAS, they remain modest by GWAS standards, where thousands and sometimes millions of samples may be required to detect loci of small effect (Santure and Garant, 2018; Uffelmann et al., 2021). Consequently, our study likely had sufficient power to rule out major effect loci (Santure and Garant, 2018), but probably had limited ability to detect the small effect variants expected under a polygenic architecture. In a study similar to our own, Chow et al. (2019) reported up to 30 loci that appeared to differ between males and females in a TSD loggerhead sea turtles (\u003cem\u003eCaretta caretta\u003c/em\u003e), suggesting possible sex-specific genotypes. However, their small sample size (n = 45) and low sequencing coverage meant that none reached genome-wide significance, limiting confidence in these associations. Similarly, candidate-gene approaches have often failed to detect robust associations between genotype and sex. For instance, Matsumoto and Crews (2017) found no link between aromatase (\u003cem\u003ecyp19a1\u003c/em\u003e) polymorphisms and sex found in pond sliders, whereas \u003cem\u003eCIRBP\u003c/em\u003e variants in the common snapping turtle showed associations with sex only under specific incubation conditions (Schroeder et al., 2016). Both patterns are biologically plausible, as genetic variation that influences sex may lie not within coding regions of known sex-related genes, but rather in upstream regulatory elements that influence temperature sensitivity of these genes (Schroeder et al., 2016; Matsumoto and Crews, 2017). Both genome-wide and candidate-gene approaches converge on the conclusion that genetic influences appear to be distributed across many loci of small effect (Santure and Garant, 2018; Chow et al., 2019), with outcomes shaped by interactions between genotype and environment (Roush and Rhen, 2018). Future progress in identifying specific sex-associated loci will likely require large scale GWAS designs involving thousands of individuals, and may necessitate multiple populations to achieve these large sample sizes, where statistical power is sufficient to detect subtle effects and to test whether associations are consistent or population-specific (Santure and Garant, 2018).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOur GBLUP analyses revealed limited predictive power for sex. Prediction accuracy in adults was significantly greater than random expectation (\u0026asymp;0.76), but AUC (\u0026asymp;0.66) and F1 score (\u0026asymp;0.35) were not different than random and suggest limited ability to predict sex. High prediction accuracy and low F1 score in adults likely occurred because of the strong female bias in the wild adults, which reduced recall of males and depressed the harmonic mean of precision and recall. While our heritability data suggest a genetic signal is present, patterns observed in our GBLUP analyses suggest that prediction accuracy may be inflated in the adult dataset. In hatchlings, prediction accuracy (\u0026asymp;0.57) and AUC (\u0026asymp;0.57) were not different from random expectations, although the F1 score (\u0026asymp;0.66) was greater than 0.5 and suggested fair discriminatory ability. Indeed, hatchlings had a more balanced sex ratio than adults, which raised the F1 score despite low overall prediction accuracy. The arc of these GBLUP results suggests, at best, that additive genomic variance contributes very weakly to sex outcomes. For comparison, prediction of binary traits with strong genetic control can reach much higher values than what we observed in turtles. In domestic dogs (\u003cem\u003eCanis lupus familiaris\u003c/em\u003e), the eye disease distichiasis has been predicted with AUC values close to 0.90, showing strong separation between affected and unaffected animals (Thorsrud et al., 2025). Root vigor in sugar beet (\u003cem\u003eBeta vulgaris\u003c/em\u003e) is controlled by only a few loci and has been predicted with AUC values near 0.98, essentially perfect classification (Biscarini, 2014). However, similar to our findings, sex in European seabass (\u003cem\u003eDicentrarchus labrax\u003c/em\u003e), a species with polygenic, temperature-driven sex determination, was predicted with only about 67% accuracy using SNP data (Palaiokostas et al., 2015). Thus, heritability estimates and some features of our GBLUP analyses suggest sex in painted turtles carries a detectable genetic signal, but for GBLUP analyses, prediction accuracy is similar to systems where the environment interacts with polygenic control.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eLong-standing theory has argued that if traits such as Tpiv are heritable, they can respond to selection imposed by changing climates, including sex-ratio selection (Janzen, 1992; Schwanz et al., 2016; Refsnider and Janzen, 2016). Empirical work among related clutches and lineages supports this prediction, documenting heritable differences in sex-ratio reaction norms among clutches and lineages (Rhen and Lang, 1998; McGaugh and Janzen, 2011; Krueger and Janzen, 2023; Krueger et al., 2025). Indeed, the present study estimated heritability of sex at \u0026asymp;0.29 in adults and \u0026asymp;0.16 in hatchlings, values similar to the effective narrow-sense heritability of \u0026asymp;0.13 for Tpiv of painted turtles in Illinois, USA (McGaugh et al. 2011). Yet, the expression of genetic variation for sex depends on whether nest temperatures fall within the TRT. Indeed, heritability estimates are often high in laboratory studies based on among-clutch (full-sibling) variance under constant incubation conditions near the Tpiv, but substantially lower when expressed at the population level under natural, fluctuating nest temperatures (McGaugh and Janzen, 2011, the present study). Family-level studies showing heritable differences in reaction norms (Rhen and Lang, 1998; McGaugh and Janzen, 2011; Krueger and Janzen, 2023), combined with our SNP-based estimates of the additive genetic variance underlying sex, nevertheless indicate that selection has something to act on. Still, several constraints limit how quickly populations may respond. For instance, much of the additive genetic variance underlying sex determination remains hidden under extreme incubation regimes. Biased adult sex ratios commonly found in TSD systems (B\u0026oacute;kony\u0026nbsp;et al. 2019) further reduce the amount of usable genetic variation by lowering effective population size and accelerating genetic drift (Refsnider and Janzen, 2016; Schwanz et al., 2020). Furthermore, the lengthy generation times exhibited by most long-lived TSD species (e.g., Sabath et al. 2016) dampen the pace of any evolutionary response to climate change. In addition, while nest-site choice has been proposed as an alternative heritable pathway for buffering sex ratios, its contribution is constrained by habitat availability and ecological context, making it a slower and less reliable avenue for adaptation than physiological traits related to Tpiv and TRT (Janzen, 1994a; Topping and Valenzuela, 2021; Crowther and Schwanz, 2025). Overall, our results indicate that heritability of traits linked to sex is modest and probably sufficient to support gradual shifts in traits linked to sexual outcomes, but it is unclear whether the pace of adaptation in these traits will be sufficient to support adaptation to rapid climate warming.\u003c/p\u003e\n\u003cp\u003eFrom a conservation perspective, our results reinforce concerns about the vulnerability of TSD species under rapid climate change. Persistent female-biased sex ratios are already observed in our Algonquin Park population at the northern-most limit of the painted turtle\u0026rsquo;s geographic range (Moldowan et al., 2020), although the cause of biased sex ratios is likely due to a warm, anthropogenic nesting environment, rather than long-term sustained warming of the environment. As temperatures continue to rise, sex-ratio bias will likely intensify, potentially pushing populations toward demographic collapse if adaptive responses cannot keep pace (Staines et al., 2023). Although the presence of additive genetic variance provides some scope for microevolution, the modest effect sizes we observed suggest that evolutionary rescue is likely to be gradual. This potential mismatch between the rate of environmental change and the strength of genetic response underscores the importance of conservation strategies that buffer populations against skewed sex ratios (Staines et al., 2023), whether through habitat protection (Rhodin et al., 2018), assisted nest management (Wnek et al., 2013), or long-term maintenance of genetic diversity (Velo-Ant\u0026oacute;n et al., 2011). Sustaining population connectivity and preserving genetic diversity may be especially important for facilitating adaptive responses under intensifying climate stress (Velo-Ant\u0026oacute;n et al., 2011; Rhodin et al., 2018). Preserving diversity not only safeguards long-term adaptive capacity but also strengthens the chance of an effective response within the short evolutionary window imposed by rapid climate warming (Janzen, 1994b; McGaugh and Janzen, 2011; Refsnider and Janzen, 2016; Crowther and Schwanz, 2025).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cem\u003eAcknowledgements\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eWe thank the EEB Department at the University of Toronto and the Algonquin Wildlife Research Station for providing accommodations and resources during fieldwork. We would like to thank the Congdon-Dickson Turtle Ecology Fund, which funded a generous portion of our genetics work for this project. We are especially grateful to Asher Cutter, Jacqueline Sztepanacz, and John Stinchcombe for their intellectual input during project design. Jessica Leivesley and Mariel Terebiznik also shared valuable expertise for SNP filtering in the early stages of this project and training in identifying sex in hatchling turtles. We would like to thank the thoughtful feedback of Rebecca Schalkowski and Phillip Pearson on an early draft of this manuscript. In addition, we greatly appreciate the field assistance of Lilian Chan and Claire Voss during the turtle nesting season at Algonquin Provincial Park. We thank Diversity Arrays Technology and their team for conducting the genotyping of our blood samples for this project. Lastly, we acknowledge AI-assisted tools (ChatGPT, OpenAI) that were used to assist with debugging, troubleshooting, and organizing code in R software. Funding was provided by NSERC Discovery grants to J.D.L. and N.R.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStatement of Authorship\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eRL and NR developed hypothesis and experimental design. RL conducted the fieldwork and laboratory experiment. RL performed the initial sex determinations for specimens with a secondary sex determination opinion from MM. RL performed the data management and statistical analyses with suggestions from JR. NR \u0026amp; JL funded the study and co-manage the painted turtle long-term data collection at the Algonquin Wildlife Research Station. RL and NR led the writing of the manuscript with contributions by all authors.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eRe\u003c/em\u003e\u003cem\u003esearch Involving Animals and their Data or Biological Material\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAnimal collecting and use was authorized a Wildlife Scientific Collectors Authorization #1100425 issued by the Ministry of Northern Development, Mines, Natural Resources and Forestry, and an Animal Use Protocol #20011948 approved and issued by the University of Toronto Local Animal Care Committee.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eAnderson CA, Pettersson FH, Clarke GM et al. (2010) Data quality control in genetic case-control association studies. Nat Protoc 5:1564\u0026ndash;1573. https://doi.org/10.1038/nprot.2010.116\u003c/li\u003e\n \u003cli\u003eAppadurai V, Bybjerg-Grauholm J, Krebs MD, Rosengren A, Buil A, Ingason A et al. (2023) Accuracy of haplotype estimation and whole genome imputation affects complex trait analyses in complex biobanks. Commun Biol 6:101\u003c/li\u003e\n \u003cli\u003eBachtrog D, Mank JE, Peichel CL, Kirkpatrick M, Otto SP, Ashman TL et al. (2014) Sex determination: why so many ways of doing it? PLoS Biol 12:e1001899\u003c/li\u003e\n \u003cli\u003eBiscarini F, Stevanato P, Broccanello C, Stella A, Saccomani M (2014) Genome-enabled predictions for binomial traits in sugar beet populations. BMC Genet 15:87\u003c/li\u003e\n \u003cli\u003eB\u0026oacute;kony V, Milne G, Pipoly I, Sz\u0026eacute;kely T, Liker A (2019) Sex ratios and bimaturism differ between temperature-dependent and genetic sex-determination systems in reptiles. BMC Evol Biol 19:57\u003c/li\u003e\n \u003cli\u003eBowden RM, Ewert MA, Nelson CE (2000) Environmental sex determination in a reptile varies seasonally and with yolk hormones. Proc Biol Sci 267:1745\u0026ndash;1749\u003c/li\u003e\n \u003cli\u003eBull JJ (1980) Sex determination in reptiles. Q Rev Biol 55:3\u0026ndash;21\u003c/li\u003e\n \u003cli\u003eBull JJ, Charnov E (1988) How fundamental are Fisherian sex ratios? [still needs journal, volume, and page range]\u003c/li\u003e\n \u003cli\u003eBull JJ, Charnov EL (1989) Enigmatic reptilian sex ratios. Evolution 43:1561\u0026ndash;1566\u003c/li\u003e\n \u003cli\u003eBull JJ, Vogt RC, McCoy CJ (1982) Sex determining temperatures in turtles: a geographic comparison. Evolution 36:326\u0026ndash;332\u003c/li\u003e\n \u003cli\u003eCarter AW, Bowden RM, Paitz RT (2017) Seasonal shifts in sex ratios are mediated by maternal effects and fluctuating incubation temperatures. Funct Ecol 31:876\u0026ndash;884\u003c/li\u003e\n \u003cli\u003eCharnov EL, Bull J (1977) When is sex environmentally determined? Nature 266:828\u0026ndash;830\u003c/li\u003e\n \u003cli\u003eChen B, Cole JW, Grond-Ginsbach C (2017) Departure from Hardy Weinberg equilibrium and genotyping error. Front Genet 8:300354\u003c/li\u003e\n \u003cli\u003eChen H, Wang C, Conomos MP, Stilp AM, Li Z, Sofer T et al. (2016) Control for population structure and relatedness for binary traits in genetic association studies via logistic mixed models. Am J Hum Genet 98:653\u0026ndash;666\u003c/li\u003e\n \u003cli\u003eChow JC, Anderson PE, Shedlock AM (2019) Sea turtle population genomic discovery: global and locus-specific signatures of polymorphism, selection, and adaptive potential. Genome Biol Evol 11:2797\u0026ndash;2806\u003c/li\u003e\n \u003cli\u003eCrews D, Wibbels T, Gutzke WHN (1989) Action of sex steroid hormones on temperature-induced sex determination in the snapping turtle (Chelydra serpentina). Gen Comp Endocrinol 76:159\u0026ndash;166\u003c/li\u003e\n \u003cli\u003eCrowther C, Schwanz LE (2025) Behavioural vs. physiological adaptation: which contributes more to the evolution of complex traits in a warming climate? J Evol Biol 38:467\u0026ndash;480\u003c/li\u003e\n \u003cli\u003eDekkers JC, Su H, Cheng J (2021) Predicting the accuracy of genomic predictions. Genet Sel Evol 53:1\u0026ndash;23\u003c/li\u003e\n \u003cli\u003eDempster ER, Lerner IM (1950) Heritability of threshold characters. Genetics 35:212\u003c/li\u003e\n \u003cli\u003eEwert MA, Etchberger CR, Nelson CE (2004) Turtle sex-determining modes and TSD patterns, and some TSD pattern correlates. In: Valenzuela N, Lance VA (eds) Temperature-dependent sex determination in vertebrates. Smithsonian Books, Washington, DC, pp 21\u0026ndash;32\u003c/li\u003e\n \u003cli\u003eEwert MA, Lang JW, Nelson CE (2005) Geographic variation in the pattern of temperature-dependent sex determination in the American snapping turtle (Chelydra serpentina). J Zool 265:81\u0026ndash;95\u003c/li\u003e\n \u003cli\u003eFalconer DS (1996) Introduction to quantitative genetics. Pearson Education India, New Delhi\u003c/li\u003e\n \u003cli\u003eFisher RA (1930) The genetical theory of natural selection. Clarendon Press, Oxford\u003c/li\u003e\n \u003cli\u003eGianola D, Van Kaam JB (2008) Reproducing kernel Hilbert spaces regression methods for genomic assisted prediction of quantitative traits. Genetics 178:2289\u0026ndash;2303\u003c/li\u003e\n \u003cli\u003eGruber B, Unmack P, Berry O, Georges A (2019) Introduction to dartR. User Manual 51:1\u0026ndash;41\u003c/li\u003e\n \u003cli\u003eHadfield JD (2008) Estimating evolutionary parameters when viability selection is operating. Proc R Soc B 275:723\u0026ndash;734\u003c/li\u003e\n \u003cli\u003eHolland JB, Nyquist WE, Cervantes-Mart\u0026iacute;nez CT, Janick J (2003) Estimating and interpreting heritability for plant breeding: an update. Plant Breed Rev 22:9\u0026ndash;112\u003c/li\u003e\n \u003cli\u003eHughes EJ, Brooks RJ (2006) The good mother: does nest-site selection constitute parental investment in turtles? Can J Zool 84:1545\u0026ndash;1554\u003c/li\u003e\n \u003cli\u003eHughes EJ (2011) The effect of sex ratio on male reproductive success in painted turtles, Chrysemys picta. Dissertation, University of Guelph\u003c/li\u003e\n \u003cli\u003eJanzen FJ, Paukstis GL (1991) Environmental sex determination in reptiles: ecology, evolution, and experimental design. Q Rev Biol 66:149\u0026ndash;179\u003c/li\u003e\n \u003cli\u003eJanzen FJ (1994a) Vegetational cover predicts the sex ratio of hatchling turtles in natural nests. Ecology 75:1593\u0026ndash;1599\u003c/li\u003e\n \u003cli\u003eJanzen FJ (1994b) Climate change and temperature-dependent sex determination in reptiles. Proc Natl Acad Sci USA 91:7487\u0026ndash;7490\u003c/li\u003e\n \u003cli\u003eJanzen FJ (1992) Heritable variation for sex ratio under environmental sex determination in the common snapping turtle (Chelydra serpentina). Genetics 131:155\u0026ndash;161\u003c/li\u003e\n \u003cli\u003eJanzen FJ, Phillips PC (2006) Exploring the evolution of environmental sex determination, especially in reptiles. J Evol Biol 19:1775\u0026ndash;1784\u003c/li\u003e\n \u003cli\u003eKrueger CJ, Janzen FJ (2023) On the origin of patterns of temperature-dependent sex determination. Evolution 77:1091\u0026ndash;1100\u003c/li\u003e\n \u003cli\u003eKrueger CJ, Girondot M, Janzen FJ (2025) The tortoise and the air: climate shapes sex-ratio reaction norm variation in turtles. Evolution qpaf126\u003c/li\u003e\n \u003cli\u003eLi S, Xu Z, Luo L, Ping J, Zhou H, Xie L, Zhang Y (2022) Latitudinal variation in the pattern of temperature-dependent sex determination in the Japanese gecko, Gekko japonicus. Animals 12:942\u003c/li\u003e\n \u003cli\u003eLockley EC, Eizaguirre C (2021) Effects of global warming on species with temperature-dependent sex determination: bridging the gap between empirical research and management. Evol Appl 14:2361\u0026ndash;2377\u003c/li\u003e\n \u003cli\u003eMcGaugh SE, Schwanz LE, Bowden RM, Gonzalez JE, Janzen FJ (2010) Inheritance of nesting behaviour across natural environmental variation in a turtle with temperature-dependent sex determination. Proc R Soc B 277:1219\u0026ndash;1226\u003c/li\u003e\n \u003cli\u003eMatsumoto Y, Crews D (2017) Genetic polymorphisms in aromatase (cyp19a1) are not associated with gonadal phenotypes in red-eared slider turtle hatchlings developed at a pivotal temperature. Sex Dev 11:151\u0026ndash;160. https://doi.org/10.1159/000471940\u003c/li\u003e\n \u003cli\u003eMcGaugh SE, Janzen FJ (2011) Effective heritability of targets of sex-ratio selection under environmental sex determination. J Evol Biol 24:784\u0026ndash;794\u003c/li\u003e\n \u003cli\u003eMcGaugh S, Bowden R, Kuo CH, Janzen F (2011) Field-measured heritability of the threshold for sex determination in a turtle with temperature-dependent sex determination. Evol Ecol Res 13:75\u0026ndash;90\u003c/li\u003e\n \u003cli\u003eMelville J, Haines ML, Boysen K, Hodkinson L, Kilian A, Smith Date KL et al. (2017) Identifying hybridization and admixture using SNPs: application of the DArTseq platform in phylogeographic research on vertebrates. R Soc Open Sci 4:161061\u003c/li\u003e\n \u003cli\u003eMitchell NJ, Janzen FJ (2010) Temperature-dependent sex determination and contemporary climate change. Sex Dev 4:129\u0026ndash;140\u003c/li\u003e\n \u003cli\u003eMittell EA, Pemberton JM, Kruuk LEB, Morrissey MB (2025) Unmeasured prior viability selection resolves the paradox of stasis for body size in wild Soay sheep. Proc Natl Acad Sci USA 122:e2513969122\u003c/li\u003e\n \u003cli\u003eMrosovsky N, Pieau C (1991) Transitional range of temperature, pivotal temperatures and thermosensitive stages for sex determination in reptiles. Amphibia-Reptilia 12:169\u0026ndash;179\u003c/li\u003e\n \u003cli\u003eMrosovsky N, Yntema CL (1980) Temperature dependence of sexual differentiation in sea turtles: implications for conservation practices. Biol Conserv 18:271\u0026ndash;280\u003c/li\u003e\n \u003cli\u003eO\u0026rsquo;Steen S, Janzen FJ (1999) Embryonic temperature affects metabolic compensation and thyroid hormones in hatchling snapping turtles. Physiol Biochem Zool 72:520\u0026ndash;533\u003c/li\u003e\n \u003cli\u003ePalaiokostas C, Bekaert M, Taggart JB, Gharbi K, McAndrew BJ, Chatain B et al. (2015) A new SNP-based vision of the genetics of sex determination in European sea bass (Dicentrarchus labrax). Genet Sel Evol 47:68\u003c/li\u003e\n \u003cli\u003ePearse DE, Janzen FJ, Avise JC (2002) Multiple paternity, sperm storage, and reproductive success of female and male painted turtles (Chrysemys picta) in nature. Behav Ecol Sociobiol 51:164\u0026ndash;171\u003c/li\u003e\n \u003cli\u003ePen I, Uller T, Feldmeyer B, Harts A, While GM, Wapstra E (2010) Climate-driven population divergence in sex-determining systems. Nature 468:436\u0026ndash;438\u003c/li\u003e\n \u003cli\u003eP\u0026eacute;rez P, de Los Campos G (2014) Genome-wide regression and prediction with the BGLR statistical package. Genetics 198:483\u0026ndash;495\u003c/li\u003e\n \u003cli\u003eRage JC (1998) Latest Cretaceous extinctions and environmental sex determination in reptiles. Bull Soc G\u0026eacute;ol Fr 169:479\u0026ndash;483\u003c/li\u003e\n \u003cli\u003eRefsnider JM, Milne-Zelman C, Warner DA, Janzen FJ (2014) Population sex ratios under differing local climates in a reptile with environmental sex determination. Evol Ecol 28:977\u0026ndash;989\u003c/li\u003e\n \u003cli\u003eRefsnider JM, Janzen FJ (2016) Temperature-dependent sex determination under rapid anthropogenic environmental change: evolution at a turtle\u0026rsquo;s pace? J Hered 107:61\u0026ndash;70\u003c/li\u003e\n \u003cli\u003eRhen T, Lang JW (1995) Phenotypic plasticity for growth in the common snapping turtle: effects of incubation temperature, clutch, and their interaction. Am Nat 146:726\u0026ndash;747\u003c/li\u003e\n \u003cli\u003eRhen T, Lang JW (1998) Among-family variation for environmental sex determination in reptiles. Evolution 52:1514\u0026ndash;1520\u003c/li\u003e\n \u003cli\u003eRhen T, Metzger K, Schroeder A, Woodward R (2007) Expression of putative sex-determining genes during the thermosensitive period of gonad development in the snapping turtle, Chelydra serpentina. Sex Dev 1:255\u0026ndash;270\u003c/li\u003e\n \u003cli\u003eRhodin AGJ, Stanford CB, Van Dijk PP, Eisemberg C, Luiselli L, Mittermeier RA et al. (2018) Global conservation status of turtles and tortoises (order Testudines). Chelonian Conserv Biol 17:135\u0026ndash;161\u003c/li\u003e\n \u003cli\u003eRao CR (1948) Large sample tests of statistical hypotheses concerning several parameters with applications to problems of estimation. Math Proc Cambridge Philos Soc 44:50\u0026ndash;57\u003c/li\u003e\n \u003cli\u003eRopp AJ, Reece KS, Snyder RA, Song J, Biesack EE, McDowell JR (2023) Fine-scale population structure of the northern hard clam (Mercenaria mercenaria) revealed by genome-wide SNP markers. Evol Appl 16:1422\u0026ndash;1437\u003c/li\u003e\n \u003cli\u003eRoush D, Rhen T (2018) Developmental plasticity in reptiles: critical evaluation of the evidence for genetic and maternal effects on temperature-dependent sex determination. J Exp Zool A Ecol Integr Physiol 329:287\u0026ndash;297\u003c/li\u003e\n \u003cli\u003eSabath N, Itescu Y, Feldman A, Meiri S, Mayrose I, Valenzuela N (2016) Sex determination, longevity, and the birth and death of reptilian species. Ecol Evol 6:5207\u0026ndash;5220\u003c/li\u003e\n \u003cli\u003eSaito T, Rehmsmeier M (2015) The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS One 10:e0118432\u003c/li\u003e\n \u003cli\u003eSamson J (2003) The life history strategy of a northern population of midland painted turtle, Chrysemys picta marginata. Dissertation, University of Guelph\u003c/li\u003e\n \u003cli\u003eSantidri\u0026aacute;n Tomillo P (2022) When population-advantageous primary sex ratios are female-biased: changing concepts to facilitate climate change management in sea turtles. Clim Change 175:15\u003c/li\u003e\n \u003cli\u003eSanture AW, Garant D (2018) Wild GWAS\u0026mdash;association mapping in natural populations. Mol Ecol Resour 18:729\u0026ndash;738\u003c/li\u003e\n \u003cli\u003eSarre SD, Georges A, Quinn A (2004) The ends of a continuum: genetic and temperature-dependent sex determination in reptiles. BioEssays 26:639\u0026ndash;645\u003c/li\u003e\n \u003cli\u003eSilber S, Geisler JH, Bolortsetseg M (2011) Unexpected resilience of species with temperature-dependent sex determination at the Cretaceous\u0026ndash;Palaeogene boundary. Biol Lett 7:295\u0026ndash;298\u003c/li\u003e\n \u003cli\u003eSchroeder AL, Metzger KJ, Miller A, Rhen T (2016) A novel candidate gene for temperature-dependent sex determination in the common snapping turtle. Genetics 203:557\u0026ndash;571. https://doi.org/10.1534/genetics.115.182840\u003c/li\u003e\n \u003cli\u003eSchwanz LE, Cordero GA, Charnov EL, Janzen FJ (2016) Sex-specific survival to maturity and the evolution of environmental sex determination. Evolution 70:329\u0026ndash;341\u003c/li\u003e\n \u003cli\u003eSchwanz LE, Janzen FJ, Proulx SR (2010) Sex allocation based on relative and absolute condition. Evolution 64:1331\u0026ndash;1345\u003c/li\u003e\n \u003cli\u003eSchwanz LE, Georges A, Holleley CE, Sarre SD (2020) Climate change, sex reversal and lability of sex-determining systems. J Evol Biol 33:270\u0026ndash;281\u003c/li\u003e\n \u003cli\u003eSchwanz LE, Georges A (2021) Sexual development and the environment: conclusions from 40 years of theory. Sex Dev 15:7\u0026ndash;22\u003c/li\u003e\n \u003cli\u003eSchwarzkopf L, Brooks RJ (1985) Sex determination in northern painted turtles: effect of incubation at constant and fluctuating temperatures. Can J Zool 63:2543\u0026ndash;2547\u003c/li\u003e\n \u003cli\u003eSchwarzkopf L, Brooks RJ (1987) Nest-site selection and offspring sex ratio in painted turtles, Chrysemys picta. Copeia 1987:53\u0026ndash;61\u003c/li\u003e\n \u003cli\u003eShaffer HB, Minx P, Warren DE, Shedlock AM, Thomson RC, Valenzuela N et al. (2013) The western painted turtle genome, a model for the evolution of extreme physiological adaptations in a slowly evolving lineage. Genome Biol 14:R28\u003c/li\u003e\n \u003cli\u003eSouza CSD, Santos VSD, Martins Filho S (2024) Genomic prediction using the lmekin function from the coxme R package. Acta Sci Agron 46:e64243\u003c/li\u003e\n \u003cli\u003eSrivastava AK, Williams SM, Zhang G (2023) Heritability estimation approaches utilizing genome-wide data. Curr Protoc 3:e734\u003c/li\u003e\n \u003cli\u003eStaines MN, Versace H, Lalo\u0026euml; JO, Smith CE, Madden Hof CA, Booth DT et al. (2023) Short-term resilience to climate-induced temperature increases for equatorial sea turtle populations. Glob Change Biol 29:6546\u0026ndash;6557\u003c/li\u003e\n \u003cli\u003eTelemeco RS, Gangloff EJ, Cordero GA, Mitchell TS, Bodensteiner BL, Holden KG et al. (2016) Reptile embryos lack the opportunity to thermoregulate by moving within the egg. Am Nat 188:E13\u0026ndash;E27\u003c/li\u003e\n \u003cli\u003eTh\u0026eacute;pot D (2021) Sex chromosomes and master sex-determining genes in turtles and other reptiles. Genes 12:1822\u003c/li\u003e\n \u003cli\u003eThorsrud JA, Evans KM, Quigley KC, Srikanth K, Huson HJ (2025) Performance comparison of genomic best linear unbiased prediction and four machine learning models for estimating genomic breeding values in working dogs. Animals 15:408\u003c/li\u003e\n \u003cli\u003eTopping NE, Valenzuela N (2021) Turtle nest-site choice, anthropogenic challenges, and evolutionary potential for adaptation. Front Ecol Evol 9:808621\u003c/li\u003e\n \u003cli\u003eTurner SD (2014) qqman: an R package for visualizing GWAS results using QQ and manhattan plots. bioRxiv 005165\u003c/li\u003e\n \u003cli\u003eUffelmann E, Huang QQ, Munung NS, De Vries J, Okada Y, Martin AR et al. (2021) Genome-wide association studies. Nat Rev Methods Primers 1:59\u003c/li\u003e\n \u003cli\u003eVanRaden PM (2008) Efficient methods to compute genomic predictions. J Dairy Sci 91:4414\u0026ndash;4423\u003c/li\u003e\n \u003cli\u003eVelo-Ant\u0026oacute;n G, Becker CG, Cordero-Rivera A (2011) Turtle carapace anomalies: the roles of genetic diversity and environment. PLoS One 6:e18714\u003c/li\u003e\n \u003cli\u003eWang J, Zhou Z, Zhang Z, Li H, Liu D, Zhang Q et al. (2018) Expanding the BLUP alphabet for genomic prediction adaptable to the genetic architectures of complex traits. Heredity 121:648\u0026ndash;662\u003c/li\u003e\n \u003cli\u003eWarner DA, Shine R (2008) The adaptive significance of temperature-dependent sex determination in a reptile. Nature 451:566\u0026ndash;568\u003c/li\u003e\n \u003cli\u003eWickham H (2011) ggplot2. Wiley Interdiscip Rev Comput Stat 3:180\u0026ndash;185\u003c/li\u003e\n \u003cli\u003eWnek JP, Bien WF, Avery HW (2013) Artificial nesting habitats as a conservation strategy for turtle populations experiencing global change. Integr Zool 8:209\u0026ndash;221\u003c/li\u003e\n \u003cli\u003eWray NR, Yang J, Goddard ME, Visscher PM (2010) The genetic interpretation of area under the ROC curve in genomic profiling. PLoS Genet 6:e1000864\u003c/li\u003e\n \u003cli\u003eYang J, Weedon MN, Purcell S, Lettre G, Estrada K, Willer CJ et al. (2011) Genomic inflation factors under polygenic inheritance. Eur J Hum Genet 19:807\u0026ndash;812\u003c/li\u003e\n \u003cli\u003eYarberry W (2021) Dplyr. In: CRAN recipes: DPLYR, stringr, lubridate, and regex in R. Apress, Berkeley, CA, pp 1\u0026ndash;58\u003c/li\u003e\n \u003cli\u003eYntema CL (1979) Temperature levels and periods of sex determination during incubation of eggs of Chelydra serpentina. J Morphol 159:17\u0026ndash;27\u003c/li\u003e\n \u003cli\u003eYntema CL, Mrosovsky N (1980) Sexual differentiation in hatchling loggerheads (Caretta caretta) incubated at different controlled temperatures. Herpetologica 36:33\u0026ndash;36\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"heredity","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"hdy","sideBox":"Learn more about [Heredity](http://www.nature.com/hdy/)","snPcode":"41437","submissionUrl":"https://mts-hdy.nature.com/cgi-bin/main.plex","title":"Heredity","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9324153/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9324153/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn species with temperature-dependent sex determination (TSD), incubation temperature during embryonic development determines sex. Because sex ratios influence population stability, environmental control of sex ratio makes populations vulnerable to warming, where increases in incubation temperature can result in sex ratio skew. The capacity of populations with TSD to adapt to warmer thermal regimes depends on whether traits related to the thermal sensitivity of sexual outcomes have a genetic basis. Here, we explore the genetic basis of sex in a species with TSD. We use 271 hatchling painted turtles (\u003cem\u003eChrysemys picta\u003c/em\u003e) reared under controlled conditions at a temperature that produces both sexes, and we couple these data with 474 wild adult turtles. We used genome-wide association (GWAS) to test for genotype–sex associations and genomic best linear unbiased prediction (GBLUP) analyses to assess the predictive power of genome-wide single-nucleotide polymorphisms. GWAS suggested that no individual loci were strongly associated with sex in hatchlings or adults, and GBLUP analyses revealed limited predictive power for sex in adults and hatchlings. However, heritability was ≈0.29 in adults and ≈0.16 in hatchlings, indicating that additive genetic variance explains a measurable portion of variation in sex among lab-reared and wild turtles. Together, these results indicate that sex in TSD turtles can be influenced by many small-effect genetic variants, with a modest additive genetic effect. Our findings demonstrate that while sex in TSD species is environmentally driven, sex also carries a genetic basis that could support adaptation in the face of changing climate.\u003c/p\u003e","manuscriptTitle":"Genetic influence on the sex ratio of a turtle with temperature-dependent sex determination","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-30 13:04:52","doi":"10.21203/rs.3.rs-9324153/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-12T21:09:44+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2026-04-22T13:56:57+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2026-04-22T12:42:38+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2026-04-22T07:43:56+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-05T05:05:29+00:00","index":"","fulltext":""},{"type":"submitted","content":"Heredity","date":"2026-04-05T05:05:28+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"heredity","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"hdy","sideBox":"Learn more about [Heredity](http://www.nature.com/hdy/)","snPcode":"41437","submissionUrl":"https://mts-hdy.nature.com/cgi-bin/main.plex","title":"Heredity","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"00f0688b-9fac-410e-980c-a6802651622c","owner":[],"postedDate":"April 30th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-12T21:09:44+00:00","index":1,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":65738221,"name":"Biological sciences/Genetics/Genomics/Conservation genomics"},{"id":65738222,"name":"Biological sciences/Evolution/Evolutionary genetics"}],"tags":[],"updatedAt":"2026-04-30T13:04:52+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-30 13:04:52","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9324153","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9324153","identity":"rs-9324153","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}