Detecting Variants near repeats

Question

[Adrian Pelin]

Hello,

I am using FreeBayes to quantify variation in a repetitive eukaryote. Normally when this is done, the analysis is focused on single copy regions, to avoid quantifying variation in repetitive areas.

What I found from past experience is that even after filtering out variants above a certain coverage, variants near repeats tend to always be false positives. Is there a way to exclude the analysis from areas near repeats?

Adrian

[Adrian Pelin]

I suppose I would have to create a repeat file myself, since I am working on a non model fungus that I just sequenced. Is there a way to intersect a VCF file with bedtools to remove SNPs close to annotated repeats?

Adrian

[Ashutosh Pandey]

Yes. Bedtools supports multiple formats. Checks this:http://bedtools.readthedocs.org/en/latest/content/tools/intersect.html

Answer 1

[Ashutosh Pandey]

You can download bed files for various repeats from here.

Select your organism of interest and select "Repeats" in group. You can then combine multiple repeat bed files together using bedtools.

Then you can use bedtools intersect to find overlapping region between your SNP file (in some format compatible with bed tools) and the repeats bed file.

[Matteo Schiavinato]

I stumbled on this post because I am struggling with some repetitive regions containing SNPs that bcftools calls. Given that obviously the tool cannot distinguish between a repetitive region and not, unless you tell it to do so, is it common practice to literally trim out SNPs in repetitive regions after calling? Or people usually align DNA reads on hard masked DNA reference sequences to avoid even having mapping results there?