How can I detect false positive deletion in the output of Pindel tool?

Question

[Medhat]

How to detect false positive deletions in the output of pindel tool?

I am using Pindel to have information about large deletion in a genome, my approach to extract the information line from the "some_file_name-p_D" and sort the the extracted file by the support column "which gives the number of reads that support the hypotheses of deletion" but I still have a lot of false positive results.

Any suggestion or better approach?

Answer 1

[poisonAlien]

You can take a look at filtering parameters explained in varscan2 paper, to reduce potential false positives.

[Medhat]

You are talking about using totally different program varscan2 instead of Pindel am I right?

[poisonAlien]

No. You can still use the Pindel output. Those filters explained in the paper can be applied to any output. They are just potential sources of false positives, like variant lying in extreme ends of reads, variants with high mismatch quality, strand bias etc.