a realiable algorithm that calls INDELS of length 50bp-300bp in cancer genomes

Question

[Bogdan]

Dear all,

please would you advise, based on your experience, which algorithm is more reliable in confidently calling INDELS in cancer genomes (I am referring especially to INDELS of length > 50bp, let's say 50bp-500bp long).

many thanks,

-- bogdan

[Bogdan]

Thank you Kevin. I have been inspecting cgpPINDEL : https://github.com/cancerit/cgpPindel, as they provide also a docker container. Please would you let me know : is cgpPINDEL largely equivalent to PINDEL ? Thanks !

[Kevin Blighe]

Replied above!

Answer 1

[Kevin Blighe]

pindel can do it but, when you get near to the 500bp mark, are you not then in the realm of small somatic copy number alterations? BBMap can also do this: detect large indels

In all cases, read length and insert size are obviously key.

Kevin

[Kevin Blighe]

Yes, cgpPindel just appears to be wrapper of Pinel. Looks good.

[Bogdan]

thanks a lot Kevin ! may I add, have you used it extensively ? does it require a lot of filtering of the VCF files in order to get the reliable variants ?

Answer 2

[Eric T.]

LUMPY and Manta do well in this range, provided you have sequencing coverage around the relevant breakpoints.

[Bogdan]

thank you Eric. good to hear from you ! please would you have any comments about DELLY, how well it does to call the INDELs in 30-300 bp range ? thanks.