Window size vs targets - using targets only?

[ranasats]

Hi,

We have a pipeline utilizing ABRA and want to update it to incorporate ABRA2. We have a list of indel realignment targets and use that along with an array of k-mers for realignment in our current ABRA workflow.

We want to continue using our input list. Is it necessary to use window-size in this case, or can the window size be automatically determined as the length of each target in our bed? Or rather, should we provide the window size as the largest size of our targets?

Thanks

[mozack]

No need to specify window-size here. The targets will be used. Target regions > 400nt will be split internally.

[ranasats]

No need to specify window-size here. The targets will be used. Target regions > 400nt will be split internally.

Thanks for your reply! We have tried doing this, but it is leading us to miss larger deletions and insertions. We have a few examples of ~1000, ~1200, and ~1500 bp deletions being misaligned by ABRA2 when they were aligned properly by ABRA. We are using standard 100bp paired-end illumina sequencing. Is it worthwhile in this case to increase to a larger window size to allow for the capture of these events?

[mozack]

Ah, I see. Yes, you'll need to experiment with window size for those larger scale deletions.