DELLY produces empty output

Question

[liorglic]

Hello,
I am trying DELLY v0.8.2 for the first time, using ~x40 coverage data on a plant genome. The bam file was obtained by running bwa mem, then sorting and indexing the result.
I am using default params for DELLY. My run results in a bcf file with no data (i.e. header is present, but no variants detected) and prints the following message to stdout:

Sample:SRR1945819_map_to_pan.sort,ReadSize=51,Median=407,MAD=11,UniqueDiscordantPairs=230596

I find it very unlikely that there are no SVs between my sample and the reference, so I guess something is wrong with my data or the way I use DELLY.
Any ideas on what might be going wrong?
Thanks!

[Kevin Blighe]

You may consider posting on the GitHub issues page: https://github.com/dellytools/delly/issues

If you take that option, link back to this thread.

[German.M.Demidov]

DELLY has a restriction on mapping quality - and your reads are 51bp (if I am correct), so some problems with multi-mappers and thus "hidden" SVs are expected, but surely not empty output. I would try other callers such as Manta and if nothing works, write to the author of DELLY.

[liorglic]

Tanks @German.M.Demidov - I am getting the same result with another bam file which has lower depth (~x10) but longer reads (101). Interestingly, I've just tried Manta on the same data and got an empty vcf again! I am starting to wonder what's wrong with this data. Perhaps PE/SR approaches are incapable of dealing with them for some reason. Any idea on what to try next?

Answer 1

[liorglic]

OK, this is a bit embarrassing, but turns out the problem was not with DELLY (or Manta), but rather with bcftools. Turns out I was using a very old version (0.1.19) that apparently can't read the new bcf. Once upgraded to 1.10.2, I am now able to read the output bcf, which is not empty at all.