Menu

#113 prophet BLOSUM query

open
5
2008-12-16
2008-04-25
Anonymous
No

Dear EMBOSS users,

I am new to protein profile searches and am using prophecy and prophet
for a distant similarity search. I'm trying to use BLOSUM45, but prophet
is reporting BLOSUM62.

I'm generating profiles using prophecy and the BLOSUM45 matrix, thus:

prophecy -auto -warning -type H -datafile EBLOSUM45 -name
homologene_1825.aln -outfile homologene_1825.prophecy

This seems fine.

I'm using the profiles to search a protein database, thus:

prophet -auto -warning -infile homologene_1825.prophecy -outfile
prophet_out.txt -sequence
NC_000909_Methanocaldococcus_jannaschii_DSM_2661_prots.fasta

I am puzzled by this line I get in my prophet output (prophet_out.txt in
this case):

# Matrix: EBLOSUM62

I can't seem to make it go away. The profile file itself
(homologene_1825.prophecy) does correctly report BLOSUM45:

Matrix EBLOSUM45

Is there any way to make prophet also use BLOSUM45? (Does it in fact
make any difference to prophet?)

Thank you very much for advice.

Best regards,

Discussion

  • Peter Rice

    Peter Rice - 2008-12-16

    This is the align output needing a matrix to calculate the statistics. We need to find a way to calculate the statistics for the profile and then to use the profile name as the matrix in the output.

    Not immediately obvious how easy this will be. Certainly setting the matrix name in prophet.c does not help - alignConsStats needs a matrix to work through and calculate

    Need to avoid calling this function in the alignWrite functions - but can we set the pairwise values correctly without it?

     
  • Peter Rice

    Peter Rice - 2008-12-16
    • assigned_to: nobody --> peterrice