Easier way to filter contaminants based on protein similarity ?

Question

[Biogeek]

Hey guys,

I'm familiar with the idea of blob plots, but for filtering off contaminants in a plant I've sequenced. I've so far used a manual screening protocol, whereby I take the top hit of each query ( sim identity , but score and e value) . I then use the uniprot and trembl species identifiers to filter off plausible contaminants such as bacteria etc.

Doing this takes a hellish long time in excel, manually. Is there anyway I can filter off all hits which are not in green plants? I am using the entire uniprot and trembl as the plant I'm working on commonly has lots of bacteria and fungi on it. I have the uniprot/ trembl identifier available to me.

Any help appreciated.

Answer 1

[Elisabeth Gasteiger]

Have you tried to perform your BLAST searches against the Plants subsection of UniProtKB in the first place? e.g. on the UniProt website http://www.uniprot.org/blast, or programmatically using one of

• EBI BLAST: http://www.ebi.ac.uk/Tools/webservices/#sequence_similarity_search_sss
• ExPASy BLAST: http://web.expasy.org/blast/ resp. http://web.expasy.org/blast/blast+_help.html#prog_access