polyRAD: Genotype calling with uncertainty from sequencing data in polyploids and diploids

Abstract

17Low or uneven read depth is a common limitation of genotyping-by-sequencing (GBS) 18 and restriction site-associated DNA sequencing (RAD-seq), resulting in high missing data rates, 19 heterozygotes miscalled as homozygotes, and uncertainty of allele copy number in heterozygous 20 polyploids. Bayesian genotype calling can mitigate these issues, but previously has only been 21 implemented in software that requires a reference genome or uses priors that may be 22 inappropriate for the population. Here we present several novel Bayesian algorithms that 23 estimate genotype posterior probabilities, all of which are implemented in a new R package, 24 polyRAD. Appropriate priors can be specified for mapping populations, populations in Hardy-25Weinberg equilibrium, or structured populations, and in each case can be informed by genotypes 26 at linked markers. The polyRAD software imports read depth from several existing pipelines, 27 and outputs continuous or discrete numerical genotypes suitable for analyses such as genome-28 wide association and genomic prediction. 29 30 Approximately 70% of vascular plant species are recent polyploids, yet genomic 31 resources and bioinformatics tools for polyploids typically lag behind those for diploids (Moghe 32 and Shiu 2014; Renny-Byfield and Wendel 2014; Bourke et al. 2018b). Reduced representation 33 DNA sequencing methods, such as genotyping-by-sequencing (GBS) and restriction site-34 associated DNA sequencing (RAD-seq), have made high-density genotyping considerably more 35 accessible and affordable (Poland and Rife 2012; Davey et al. 2013). However, the two most 36 popular pipelines for processing GBS and RAD-seq data, Stacks (Catchen et al. 2013) and 37 TASSEL (Glaubitz et al. 2014), do not output polyploid genotypes. Though pipelines for 38polyploids are available, each have limitations that prevent their general application. For 39 example, the UNEAK pipeline is designed for diploidized polyploids only (Lu et al. 2013). 40 HaploTag is specialized for self-fertilizing polyploids (Tinker et al. 2016). FreeBayes and 41 GATK can output polyploid genotypes, but require a reference genome (McKenna et al. 2010; 42 Garrison and Marth 2012). The software EBG imports read depth from other pipelines to 43 estimate auto-or allopolyploid genotypes (Blischak et al. 2018) but requires allele frequency 44 estimations from the parent species for allopolyploids. The R package updog estimates 45 polyploid genotypes from read depth, modeling preferential pairing and accounting for multiple 46 technical issues that can arise with sequencing data, and can output posterior mean genotypes 47

Expected Genotype Quality and Diploidized Marker Data from Genotyping‐by‐Sequencing of Urochloa spp. Tetraploids

“…The aforementioned considerations are appropriate for genotype calling based on the posterior mode. An alternative approach is to estimate allele dosage based on the posterior mean, which produces fractional genotype calls (Ashraf et al, 2014;Sverrisdoìttir et al, 2017;Clark et al, 2019). Such data are suitable when additive models are used in association analysis and genome-wide prediction, but a number of genetic analyses require integral estimates of dosage, including linkage analysis (Hackett et al, 2013;Zheng et al, 2016), dominance effects (Rosyara et al, 2016;Endelman et al, 2018), and haplotype inference (Su et al, 2008;Aguiar and Istrail, 2013).…”

Expected Genotype Quality and Diploidized Marker Data from Genotyping-by-Sequencing of Urochloa spp. Tetraploids

“…For GBS data, the allele signal intensity is the read count, which can be analyzed using the aforementioned classifiers, but the focus of this manuscript is genotype calling based on a binomial model. The binomial model is central to the well-established software packages GATK (McKenna et al, 2010;Depristo et al, 2011) and FreeBayes (Garrison and Marth, 2012), as well as more recent tools developed specifically for polyploids (Blischak et al 2018;Clark et al 2018;Gerard et al 2018). It is generally recognized that higher read depth is needed to estimate allele dosage in polyploids, but precise guidelines are lacking.…”

Genotyping Polyploids from Messy Sequencing Data

“…More generally, polyploidy plays a key role in plant evolution [Otto andWhitton, 2000, Soltis et al, 2014] and plant biodiversity [Soltis and Soltis, 2000], and understanding polyploidy is important when performing genomic selection and predicting important agronomic traits [Udall and Wendel, 2006]. Consequently there is strong interest in genotyping polyploid individuals, and indeed the last decade has seen considerable research into genotyping in both non-NGS data [Voorrips et al, 2011, Serang et al, 2012, Garcia et al, 2013, Bargary et al, 2014, Mollinari and Serang, 2015, Schmitz Carley et al, 2017 and NGS data [McKenna et al, 2010, Li, 2011, Garrison and Marth, 2012, Blischak et al, 2016, Maruki and Lynch, 2017, Blischak et al, 2018, Clark et al, 2018.…”