STAR alignment - splice acceptor site vs mismatch

[Matt.Kanke]

This is more a question about how STAR works. In figure 1 of the paper (here), the aligner can either look for a splice acceptor site or extend beyond the mismatch. I was wondering how this happens, does STAR try to align the second half of the read, and if that is not possible, will extend beyond the mismatch and try aligning the next part? Or is something entirely different going on?

Thanks,
Matt

[alexdobin]

This is more a question about how STAR works. In figure 1 of the paper (here), the aligner can either look for a splice acceptor site or extend beyond the mismatch. I was wondering how this happens, does STAR try to align the second half of the read, and if that is not possible, will extend beyond the mismatch and try aligning the next part? Or is something entirely different going on?

Thanks,
Matt

Hi Matt,

STAR will indeed re-start the search for maximal mapped prefixes in the unmapped portion of the read. Moreover, the search is started not just at the beginning/end of the reads, but also at internal positions in the read that are defined by --seedSearchStartLmax and --seedSearchStartLmaxOverLread parameters.

Cheers
Alex

[Matt.Kanke]

Thanks for the response.