Ambiguous base calls causing crashes

[lizardstarks]

trgt/src/trgt/genotype/consensus.rs

Line 81 in fe63314

_ => panic!("Encountered unexpected base"),

The consensus step of variant calling in TRGT requires all base calls to be unambiguous, and "N" is not permitted. Unfortunately this is causing crashes on one of my sources of data which includes Ns.

Is it possible to build in support for handling reads with ambiguous bases?

[egor-dolzhenko]

Thank you for pointing this out! What kind of data are you working with?

[lizardstarks]

Thanks for looking into it!
It's a bit of an unconventional use case, but I'm testing TRGT performance on Illumina technologies for some evaluations of unexpanded STRs. TRGT is working pretty well on Illumina data, aside from crashing on reads that contain isolated Ns in the middle of otherwise 'good' aligned sequence with high base Q.

[lizardstarks]

To explain a bit more about my goal here, I'd like to combine data we're getting from PacBio and Illumina short reads, but I'm unsure if PacBio wants to support such a use case.
In the license, it's a little unclear to me if augmenting a PacBio analysis with reads from another technology would be considered outside the terms. I'd appreciate feedback on whether this would be a supported/permitted use.

[egor-dolzhenko]

Thank you for the question. TRGT is meant to be used with long-read PacBio data only. You could use a method like Straglr that supports different kinds of data. If you'd like, I would be very happy to discuss your analysis in more detail (if so, please reach out by email).