any tutorials for structural variants filtering and prioritization

Question

[Ming Tommy Tang]

Outputs from structural variants callers like breakdancer, lumpy and delly contains false positives.

How do you filter out those false positives and then prioritize the SVs? Anyone has a good resources or experiences on that ?

Thanks!

[lh3]

Check out: http://bcb.io/2014/08/12/validated-whole-genome-structural-variation-detection-using-multiple-callers/

Personally, I find the builtin delly filter works reasonably well.

[Ming Tommy Tang]

Thanks Heng for your reply! I happen to read this post before. I am wondering have you written any tutorials for that?

Answer 1

[Chirag Nepal]

Any idea how you would filter predictions from breakdancer?

One simple way is filter the output based on prediction score and read counts

# $9= prediction score
# $10= supporting reads

cat libraryP1.ctx | awk 'BEGIN {OFS="\t"} { if ($9 >= 40 && $10 >= 5) { print $0 }}' | wc -l