Genetic diversity and sex‐biased dispersal in the brown spotted pitviper (Protobothrops mucrosquamatus): Evidence from microsatellite markers

Abstract Dispersal plays a vital role in the geographical distribution, population genetic structure, quantity dynamics, and evolution of a species. Sex‐biased dispersal is common among vertebrates and many studies have documented a tendency toward male‐biased dispersal in mammals and female‐biased dispersal in birds. However, dispersal patterns in reptiles remain poorly understood. In this study, we explored the genetic diversity and dispersal patterns of the widely distributed Asian pitviper Protobothrops mucrosquamatus. In total, 16 polymorphic microsatellite loci were screened in 150 snakes (48 males, 44 females, 58 samples without sex information) covering most of their distribution. Microsatellite analysis revealed high genetic diversity in P. mucrosquamatus. Bayesian clustering of population assignment identified two major clusters for all populations, somewhat inconsistent with the mitochondrial DNA phylogeny of P. mucrosquamatus reported in previous research. Analyses based on 92 sex‐determined and 37 samples of P. mucrosquamatus from three small sites in Sichuan, China (Mingshan, Yibin, and Zizhong) consistently suggested female‐biased dispersal in P. mucrosquamatus, which is the first example of this pattern in snakes. The female‐biased dispersal patterns in P. mucrosquamatus may be explained by local resource competition.

Protobothrops mucrosquamatus (Cantor, 1839) (Figure 1) is a medium-sized Asian pitviper distributed in southwest and southeast China, Laos, northern Bangladesh, Vietnam, northern Myanmar, and northeastern India (Zhao, 2006). Due to the wide distribution of P. mucrosquamatus, it is easy to be encountered in the field. Thus, it is a very ideal species to explore its genetics, evolution, and ecology. Zhong et al. (2017) examined and morphologically compared 142 specimens of P. mucrosquamatus and identified sexual dimorphism within the species but no significant morphological differences among the populations, despite their wide distribution. Based on two mitochondrial DNA fragments and two nuclear genes, Guo et al. (2019) explored the genetic diversity and population evolutionary history of P. mucrosquamatus and found five geographically structured and well-supported mtDNA matrilineal lineages within the species. However, due to the limited genes, the DNA sequences did not provide much additional information on population structure.
Microsatellites, also known as simple sequence repeats (SSR), are recurring motifs of 1-6 nucleotides found in the genomes of eukaryotes (Selkoe & Toonen, 2006). In comparison to other polymerase chain reaction (PCR)-based methods, including inter-simple sequence repeat (ISSR), randomly amplified polymorphic DNA (RAPD), and amplified fragment length polymorphism (AFLP), microsatellites represent a powerful marker due to their codominant inheritance and high polymorphism, and have been widely used in phylogeographic, population, and parental analyses (Guichoux et al., 2011;Hodel et al., 2016;Qin et al., 2017). In this study, based on microsatellite markers, we explored the genetic diversity and population genetic structure of P. mucrosquamatus, and determined whether sex-biased dispersal exists in this species.

| Sampling and RAD sequencing
In total, 150 P. mucrosquamatus snakes covering most of their range were collected between 1994 and 2018 through fieldwork or tissue loans from colleagues and museums ( Figure 2 and Table 1). Liver and muscle tissue samples were taken and preserved in 90% ethanol.
Whole genomic DNA was extracted using a TIANamp Genomic DNA kit (Tiangen Biotech (Beijing) Co., Ltd.) following the manufacturer's protocols.
High-quality DNA was transferred to Novogene Bioinformatics Technology Co., Ltd. for restriction site-associated DNA sequencing (RAD-seq) according to the standard protocols, in which total genomic DNA was digested with MseI restriction enzymes. The generated library was sequenced on the Illumina HiSeq 2000 platform to produce paired-end reads. The quality of the raw reads was assessed using FastQC v.0.11.9 (Brown et al., 2017). High-quality reads were clustered using CD-HIT-EST v. 4.8.1 (Li & Godzik, 2006) and assembled into contigs using Velvet v.1.2.10 (Namiki et al., 2012).

| Microsatellite amplification and genotyping
After quality filtering, the high-throughput sequencing data were screened to locate tetra-nucleotide perfect repeat microsatellite loci using MSDB v.2.4.2 software (Du et al., 2012). Primer pairs were designed using Primer v.3.0 (Untergasser et al., 2012), with amplicon size ranging from 100 to 250 bp. In total, 25 microsatellite markers were randomly selected for optimization, and 16 markers were subsequently used to evaluate the genetic diversity and dispersal patterns of P. mucrosquamatus.

| Diversity assessment
The successfully optimized microsatellites were used to evaluate the genetic diversity of P. mucrosquamatus. PCR was performed in a 25 µl volume containing 30 ng of genomic DNA, 1 µl of each primer (10 µM), 12.5 µl of 2 × T5 Super PCR Mix (PAGE) (Beijing Tsingke Biotech Co., Ltd.), and 10 µl of nuclease-free water. The cycling conditions included a hot start pre-denaturation of 95°C for 4 min, F I G U R E 1 The photo of Protobothrops mucrosquamatus in life followed by 35 cycles of denaturation at 94°C for 45 s, annealing at 61-63°C (according to each primer pair) for 30 s, extension at 72°C for 30 s, post-extension at 72°C for 10 min, and heat preservation at 10°C.
The PCR product size was measured on an ABI 3730xl DNA Analyzer (Applied Biosystems) according to each forward primer labeled with fluorescent dyes (FAM, HEX, or TAMRA) and data were obtained with GeneMapper v.4.0 (Applied Biosystems). All samples were read at least three times to reduce artificial error.
All loci were characterized, and the full dataset (150 individuals) was analyzed for various genetic diversity indices. Based on the mitochondrial DNA phylogeny of P. mucrosquamatus (Guo et al., 2019),  (Chapuis & Estoup, 2006) software to detect null alleles, stuttering, and large allele dropout errors that can occur during the interpretation of microsatellite allele sequences. If there is a higher frequency of null alleles, that is, if it exceeds 0.2 for population genetic analysis, and if it exceeds 0.08 for parental analysis, the locus can be discarded or the null allele can be eliminated by redesigning primers (Wen et al., 2013). Deviation from the Hardy-Weinberg equilibrium (HWE) was tested for each locus across and within populations by Fisher's exact test (Guo & Thompson, 1992) implemented in GenePop v.4.6 (Rousset, 2008) using a Markov chain  (Kalinowski et al., 2007). PGDSpider v.2.1.1.5 (Lischer & Excoffier, 2012) and GenAlEx v.6.5 (Peakall & Smouse, 2012) were used to perform conversions between different data formats.

| Genetic structure
STRUCTURE v.2.3.4 (Pritchard et al., 2000) was used to infer population structure and assign individuals to subpopulations following the admixture model. What is more, we use sampling location as prior (LOCPRIOR) to assist the clustering in STRUCTURE v.2.3.4.
The most likely number of genetic clusters (K) varied from K = 1 to K = 10, with a burn-in of 100,000 and MCMC repeats of 1,000,000 with 10 iterations. Results were collated using Structure Harvester v.0.6.94 (Earl & Vonholdt, 2012) and visualized using Excel. Selection of the optimal K-value was based on both the log-likelihood value closest to zero and the ΔK parameter (Evanno et al., 2005). CLUMPP v.1.1.2 (Jakobsson & Rosenberg, 2007) was used to cluster repeated sampling. Distruct v.1.1 software (Rosenberg, 2004)  among populations (F st ) were performed using GenAlEx v.6.5 (Peakall & Smouse, 2012). To delineate the major ordination pattern of P. mucrosquamatus populations, a discriminant analysis of principal components (DAPC) (Jombart et al., 2010) was performed by R v3.6.1 (R Core Team, 2019) using the adegenet package (Jombart, 2008). DAPC analysis is a multivariate method used to identify and describe clusters of genetically related individuals. Genetic variation is divided into two parts: between-group variation and within-group variation, which maximizes the former. Linear discriminants are linear combinations of alleles that best separate clusters (Deperi et al., 2018).

| Tests for sex-biased dispersal
In total, 92 sex-determined individuals (48 males, 44 females) from the SCV and SWC populations were used to evaluate sex-biased dispersal. We assessed sex-biased dispersal from three small sites We calculated F st (Hartl & Clarck, 1997), F is , genetic diversity (H s ), relatedness (r), mean assignment index (mAIc) (Favre et al., 1997), and variance of assignment index (vAIc) for each sex separately using FSTAT v.1.2. (Goudet, 1995). Statistical significance for these indices was determined by 10,000 randomizations. We chose the unbiased Weir and Cockerham estimator to calculate F st across all populations (Weir & Cockerham, 1984), with values generally higher for the philopatric sex than the dispersing sex. F is describes how well genotype frequencies within populations fit the HWE, with values generally higher for the dispersing sex than the philopatric sex. Within-group Hs values are also higher for the group with the greatest dispersal. In the case of sex-biased dispersal, mAIc values should be lower for the dispersing sex than for the philopatric sex (Lampert et al., 2003). In contrast, vAIc values should be higher for the dispersing sex because members will include both residents (with common genotypes; positive values) and immigrants (with rare genotypes; negative values). In brief, higher F is , Hs, and vAIc values and lower F st , mAIc, and r values tend to be found in the dispersing sex than in the philopatric sex .
To further verify the results of sex-biased dispersal, we analyze data from the 92 sex-determined individuals and three small sites separately, we calculated and compared relatedness values between the sexes using COANCESTRY v.1.0 with five moment and two likelihood estimators (Wang, 2011).  for total dataset are listed in Table 3.

| Population genetic structure
To analyze the genetic structure of P. mucrosquamatus populations, and SCW with SWC population were low, suggesting low genetic differentiation among them (Appendix S6). to males, but lower F st , mAIc, and r values (Table 5). When we examined the three sites separately, two out of seven relatedness indices were significantly higher in males than in females (p < .05) ( Table 6).  (Dubey et al., 2008). In addition, the mean PIC (0.879) of P. mucrosquamatus was >0.5, indicating that this species was highly genetically diverse. Higher genetic diversity could be attributed to their wide regional distribution and varied habitats.

| Genetic diversity and population structure
Based on genetic structure analysis, we detected two clusters in P. mucrosquamatus, different from previous mtDNA-based findings (Guo et al., 2019) to some extent. This difference may be due to different genetic and evolutionary patterns between mtDNA and microsatellite markers. However, these two clusters displayed significant admixture, consistent with AMOVA results, which indicated variation among individuals (Appendix S5). A standard AMOVA for the 5 populations (without a hierarchy of regions) showed that 82% of the variation was located between individuals and only 4% among populations. In China, the last global glaciation, termed the Dali glaciation (DLG), occurred during 0.07-0.01 Ma (Shi & Wang, 1979 (Shi & Wang, 1979). However, the SCV population experienced an expansion before 0.07 Ma, which may have been triggered by pre-Glacial Maximum. High temperatures.

| Sex-biased dispersal
In general, the F is , F st , r, mAIc, vAIc, and Hs parameters are indicative of sex-biased dispersal patterns. Previous studies have shown that F st is higher for the more philopatric sex than for the more dispersing sex (Goudet et al., 2002). Members of the dispersing sex also display higher F is than the philopatric sex. Furthermore, Hs is generally higher in the group showing greater dispersal. In the case of sexbiased dispersal, mAIc values are lower for the dispersing sex than for the philopatric sex (Lampert et al., 2003) (Dubey et al., 2008;Folt et al., 2019;Hofmann et al., 2012;Keogh et al., 2007;Lukoschek et al., 2008;Pernetta et al., 2011;Zwahlen et al., 2021).
However, most indices representing sex-biased dispersal did not differ significantly, which may be the result of incomplete sampling.

Several hypotheses have been proposed for female-dispersal in birds
and mammals, including local resource competition (Greenwood, 1980), local mate competition (Dobson, 1982;Perrin & Mazalov, 2000;Rivas & Burghardt, 2005), and inbreeding avoidance (Perrin & Mazalov, 2000;Pusey, 1987). Although the true mechanism of sex-biased dispersal is unknown in this species, we hypothesize local resource competition may better explain the dispersal pattern as females need to acquire more resources while avoiding increased competition for resources. P. mucrosquamatus is widely distributed in southeastern and southwestern China, Laos, Bangladesh, northern Vietnam, northern Myanmar, and northeastern India. It is one of the most widely distributed members in this genus, and its distribution covers different climates and vegetation types (Zhao, 2006).
Maybe it has something to do with the females of this species being more inclined to dispersal. help in data analysis. The editor and two anonymous reviewers are acknowledged for their invaluable comments and corrections.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All microsatellite genotypes for all individuals are deposited in Dryad ht tps://dat ad r yad.org /st ash/ share/ Ntrk9 UMZIh u7Zag 5DOv 0 c8d1y XIsF8 Fd2BJ zgGtE4WA. All genetic analyses were performed with publicly available programs.