Genotyping by Sequencing in Pooled Samples of Barley Doubled Haploid Lines: A Case Study

Genotyping by sequencing (GBS) provides an economical and rapid approach to simultaneously identify and genotype large numbers of single nucleotide polymorphisms (SNPs) in many individuals (1, 2). In a typical analysis, SNP genotyping is performed on each line individually. However, it could be useful to analyze a set of lines as a composite sample by pooling DNA. To explore the most efficient approach, we prepared a GBS library using the PstI/MspI protocol for 75 doubled haploid (DH) lines of barley and obtained 125M reads. These were used to genotype each of the lines individually and the same 125M reads were analyzed as a single, composite sample of the same 75 DH lines. We determined the number of SNPs that could be called as a function of the number of reads used (5-25M reads) and the variability surrounding the estimated allele frequencies using 1 to 5 subsets of 5-25M reads.  Using the individual analysis as a “gold standard”, we optimized the pooled analysis such that allelic frequencies were estimated correctly but using much fewer reads (15M reads). On the entire DH population, we identified the same three regions of high segregation bias by using the composite approach compared to those obtained with individual analysis. Using a composite approach, it is possible to examine 10‑fold more DH populations for an identical sequencing effort. The analyses conducted in this work likely apply to a number of other situations where the pooling of samples could prove more cost-effective than the analysis of individual samples.