GBS-SNP-CROP: a reference-optional pipeline for SNP discovery and plant germplasm characterization using variable length, paired-end genotyping-by-sequencing data

Table 2 Performance of GBS-SNP-CROP under three different sampling strategies for building the Mock Reference: Using all 48 individuals in the population (MR48), using only the 5 individuals with the highest number of parsed reads (MR05), and using only the single most read-abundant genotype (MR01)

Pipelines	Total number of centroids used to build the Mock Reference^a	Total number of paired-end reads used for SNP calling^b	Number of SNPs called^c	Avg. depth^d	Hetero (%)^e	Homo (%)^f	Missing data (%)^g	Time (hrs:mins)^h
GBS-SNP-CROP-MR48	1,276,734	92,667,123	14,712	70.74	32.47	59.31	8.20	14:30
GBS-SNP-CROP-MR05	500,795	132,920,383	20,226	71.02	34.50	57.18	8.31	12:06
GBS-SNP-CROP-MR01	229,549	154,506,669	21,318	69.34	34.51	56.85	8.29	11:03

^a Total number of non-redundant consensus sequences (centroids) identified via clustering to represent the GBS fragment space. This is also the number of FASTA entries in the “MockRef_Clusters.fasta” file
^b Number of reads retained by the pipeline after mapping procedures and thus used for SNP calling
^c Total number of SNPs called, given all SNP calling filters and genotyping criteria described in the text
^d Average read depth for all SNPs across the entire population
^e Percentage of heterozygous genotype calls
^f Percentage of homozygous genotype calls
^g Percentage of missing cells (i.e. no genotype call for a given SNP*accession combination) in the final SNP genotype matrix
^h The total computation time required for all pipeline analysis when executed on a Unix workstation with 16 GB RAM and a 2.6 GHz Dual Intel processor

ISSN: 1471-2105