Forum:Benchmark toolkits for Diploid Variant Calling

Question

[14134125465346445]

This is a post to compile all the current toolkits to evaluate/benchmark Diploid Variant Calling. Feel free to contribute as comment or editing the post.

Smash - https://github.com/amplab/smash

[Devon Ryan]

Not a toolkit, but Brad Chapman's blog contains about enough relevant comparisons that it should probably get a mention anyway.

Answer 1

[lh3]

My hapdip benchmark. It consists of two complement components: a benchmark on a pair of haploid-diploid samples (see this) and a benchmark on GIAB's high-confident regions. The former gives an unbiased global view of error rates. The latter measures how we can reduce errors in well-behaved genomic regions.