The impact of sequencing depth and DP filtering on genotyping accuracy of SNP based on next-generation sequencing.

Next-generation sequencing (NGS) is the primary method for SNP genotyping in a large-scale population. However, it still faces a challenge to balance the paradox between SNP genotyping accuracy and sequencing cost in populations characterized by high heterozygosity and low genomic linkage disequilibrium. Although sequencing depth and depth of coverage (DP) are critical parameters influencing SNP genotyping accuracy, comprehensive investigation on genotyping consistency of SNP among different sequencing depths and impact of DP on SNP genotyping accuracy of low-depth sequencing is of absence. In this study, we utilized Lueyang Black-boned chickens (n=11) to evaluate genotyping consistency of SNP at 5× and 10× sequencing depths via two strategies: (1) comparison of SNP genotyping results between 5× and 10×; (2) comparison of genotyping accuracies of 5× and 10× by employing Sanger-genotyping results of 162 SNP at the SLCO1B3 locus as the criteria. Furthermore, we investigated the effect of DP filtering on genotyping consistency between 5× and 10× by increasing DP from 0 to 5. We studied the association of genotyping results of NGS with GC contents and repeat sequences for each chromosome. The results show that the ratio of genotyping consistency is 78.07% between 5× and 10×. The ratio is 78.40% between 5× and Sanger sequencing, and increased to 87.17% for 10×. Almost all (99.18%) of inconsistent results happen in heterozygotes, of which 5×AB-10×AA/BB is mainly present in 14 microchromosomes with lengths less than 10 Mb. The 5×AB-10×AA/BB subset of inconsistent results is significantly associated with repeat sequences and GC contents. The ratio of consistent loci increases from 78.07% to 87.33% with the increasing of DP from 0 to 5. However, calling rates of SNP dramatically reduces from 100% to 35.07%. The ratio of missing loci decreases from 13.16% to 4.08%. DP filtering has no significant effect on the ratio of inconsistent loci. The results indicate that challenge of the NGS-based SNP genotyping approach focuses on genotyping of heterozygotes. Sequencing depth and sequence features of chromosomes affect genotyping accuracy. To physically increase sequencing depth improves the genotyping accuracy of SNP, whereas to algorithmically raise DP has a negligible effect. Increasing DP can significantly improve the genotyping consistency between low- and high-depth sequencing. The improving effect quickly decays with the increasing of DP, whereas calling rates of SNP dramatically reduce.