count indel by length in human reference

Question

[bioguy24]

Is there a tool or script that can calculate the ins or del length by chromosome in the ucsc.hg19.fasta (human geneome)? Thanks.

[Pierre Lindenbaum]

what 's your input ? VCF ? BAM ?

[bioguy24]

It would be a BAM comparing NS12911 or NA12878 to the hg19 reference. Is this possible? Thanks.

Answer 1

[Sean Davis]

Step 1: Call variants using a variant caller such as platypus, freebayes, samtools, or GATK.

Step 2: Write a small custom script to calculate the indel lengths per chromosome.

[bioguy24]

From a vcf I can do already, but that will only tell me what was called.... basically I am trying to get the "true" # in the human reference as compared to say NS12911 or NA12878. Is this possible? Thank you :). Edit: what I trying to do is sequence NA12878 and output the ins or del by length from the vcf. I can get that data but is there a way to get the actual or close to actual insertion/deletion by length of that known sequence?

[Sean Davis]

Well, the "true" number is not actually known, but for GIAB sample, you could get the consensus calls and use that as an approximation of the "truth".