Sunday, May 5, 2013

Blog: Blue Collar - Bioinformatics Framework for evaluating variant calling methods: comparison of aligners and callers

Here are the main points from the post, the important thing is how to use them in Paper 1.

Differences between mappers, for this we must know the true variants in the sample!

  • Extra variants, called in the evaluation data but not in the reference. These are potential false positives
  • Missing variants, found in the NA12878 reference but not in the evaluation data set. These are potential false negatives. The use of high quality reference materials from NIST allows us to be able to assess the equally important question of identify genome regions we fail to call in
  • Shared variants, called in both evaluation and reference but differently called. This could be due to allele difference, such as heterozygote versus homozygote calls or variant representation differences, such as different indel sizes.
Reason for this difference:
 Another interesting note is that: "Other critical components of assessment include comparing SNPs and indels separately and evaluating only in callable regions with 4 or more reads"

 Aligners section:
1024 of these missing variants are in regions where novoalign does not provide sufficient coverage for calling. Algorithmic changes impact low coverage regions more due to the decreased evidence and susceptibility to crossing calling coverage thresholds, so we need extra care and consideration of calls in these regions.

Notes to myself:


If I create a new merge of bams, between human cancer and mouse. And I know the variants of the human because the mapping with bwa, should I keep the number of reads of human constant? So the number of variants and quality stay the same?
For the reasons of differences, for me I would take only the first one, the other may affect in the first calling, so no change. I'd add the mix of reads form different source, and multiple alignment.

No comments:

Post a Comment