Evaluation of variant calling tools for large plant genome re-sequencing

Area of Science:

• Genomics
• Bioinformatics
• Crop Science

Background:

• Single nucleotide polymorphisms (SNPs) are crucial for genetic marker development in crops.
• Accurate SNP detection in large polyploid genomes like wheat is challenging.
• Existing variant calling methods show low concordance, and a gold standard for wheat is lacking.

Purpose of the Study:

• To evaluate seven SNP variant calling tools (FreeBayes, GATK, Platypus, Samtools/mpileup, SNVer, VarScan, VarDict).
• To compare two popular mapping tools (BWA-mem and Bowtie2).
• To assess performance on wheat whole exome capture (WEC) re-sequencing data.

Main Methods:

• Comparative analysis of seven SNP variant calling tools.
• Evaluation using two mapping tools (BWA-mem and Bowtie2).
• Testing on allohexaploid wheat WEC re-sequencing data.

Main Results:

• BWA-mem demonstrated higher mapping rate and accuracy than Bowtie2.
• Samtools/mpileup with BWA-mem mapping outperformed other pipelines in specificity and sensitivity.
• Preprocessing steps like quality trimming did not significantly impact mapping performance.

Conclusions:

• The BWA-mem and Samtools/mpileup pipeline is the optimal choice for SNP calling in wheat genome re-sequencing.
• This pipeline requires no preprocessing of raw reads before mapping.
• Findings offer guidelines for variant identification in other large polyploid crop genomes.