Adding option in --singleStrandMode to ignore transitions for X number of bases from the end of each read

[malmarri]

Hi Stephan, thank you for this great tool.

Would it be possible to add in the future an option where instead of ignoring transitions in the forward strand (C/T) and negative strand (G/A) in --singleStrandMode for the whole read, we can apply it to a specific number of bases from the end of the read set by the user? This can balance damage removal without removing a significant amount of useful data from transitions far from the end which are less likely to have damage (though do so at lower levels).

Thank you

[stschiff]

Yes, that would be very cool. However, the pileup Format as it is defined makes inferring positional information very challenging. We would have to record read starts and ends, which requires the entire algorithm to be changed to "remember" identity of reads and their position. Currently, the program is far-simpler, and simply calls per position. This allows, for example, positions to be filtered before piping them to pileupCaller.

So I think what you have in mind would arguably better be done in a separate tool outside of pileupCaller, for example in something like bamTable from our department.

I was thinking at some point to support bamTable output in pileupCaller, but have not acted on it yet.

Sorry that I cannot make this easier for now.