Realignment intervals distribution

I was just wondering what you guys thought of my realignment intervals length distribution.
This is 30Mb from a single diploid sample without prior indel position information. Approximately 60,000 events , i.e. one every fifty bases seems like a lot.
How indicative of true indels is the data from TargetCreator and IndelRealigner? Guess I'll have to check with the ug-vcf calls...
Across the genome, distribution of 'all' events is uniform.
Does multi-sample realignment improve the accuracy or efficiency of the realignment process ?

Best Answer


  • Geraldine_VdAuweraGeraldine_VdAuwera Cambridge, MAMember, Administrator, Broadie

    Not indicative at all. You shouldn't be using the intervals from RTC to assess indel distribution in your data. These just indicate regions that are a little messy in the original alignments.

  • BlueBlue Member

    Thanks for the response. I wasn't strictly intending on using the realignment intervals for inde/CNV detection, I was just wondering what they meant, and whether they could be useful for something. Would comparing the intervals between samples be informative?

    The main thing I was wondering was if the distribution of alignment in interval sizes looked reasonable to you. It's a Drosophila genome so maybe it's different from the one's you're used to.

Sign In or Register to comment.