Reader Comments

Post a new comment on this article

Comparison to VarScan?

Posted by Charles_Warden on 08 Oct 2013 at 18:46 GMT

Congrats - it looks like a lot of work went into this paper.

The paper doesn't include a comparison of SNP calls made by VarScan (which I've found to be a pretty good tool). Did you look at VarScan at all during your comparisons?

No competing interests declared.

RE: Comparison to VarScan?

liuxt005 replied to Charles_Warden on 11 Oct 2013 at 14:47 GMT

We did not look at VarScan during the comparison. The callers chosen for the comparison study were popular ones which had been used to generate the 1000 Genomes variants. Those were more popular for general NGS studies at the moment we started the project.

VarScan has two advantages: 1. Simplicity; 2. Suitable for detecting somatic mutations and copy number variations in cancer studies. “VarScan employs a robust heuristic/statistic approach to call variants that meet desired thresholds for read depth, base quality, variant allele frequency, and statistical significance” (from the VarScan web site http://varscan.sourceforg...). However, its mathematical model for single-sample SNP-calling is not deliberate as the callers included (all use probabilistic models), and it includes fewer measures for quality control than SAMtools/GATK. Furthermore, it needs pileup/mpileup files generated by SAMtools as input which need big storage space.

No competing interests declared.