I would like to compare between a set A of protein sequences (about 400 proteins) and another set B of protein sequences (about 3000 proteins). I want to see if any of the proteins in set A have a %id higher than a threshold (eg 20% or 30%) respect to any of the proteins in set B.
As the proteins in set A are trimmed previously, I would like to do glocal alignments, allowing free end-gaps in proteins in dataset B, but taking into account the entire length of the proteins in dataset A.
I am currently using Parasail in order to do pairwise alignments and obtain the %id for each comparison. However, as it is a pairwise tool, it is time-consuming to compute the comparison between each protein in dataset A to the entire dataset B, which takes about 30 minutes, where each pairwise alignment needs a little bit less than 1s to be computed.
Hence, I am looking for a BLAST-like approach, that compares protein to database. Does somebody know any? Or is there any a better way to do the comparison?
Or maybe can I modify some parameters to run BLAST taking into account a longer query length? If so, how can I do? Which parameters should be considered?