hmmsearch output multiple alignment with initial query sequences

Question

[el97004]

Hi all!

I'm running a hmmsearch using a list of proteins, and I'm wondering if the program is able to output a final multiple alignment which includes my query proteins + all significant sequences that were found in the run. I was able to output a multiple alignment of all significant hits using the -A flag:

hmmsearch -A output_msa proteins.hmm my_protein_database

but I can't find a way for outputting a final model including my query + new hits.

Thanks!

Answer 1

[Mensur Dlakic]

This can be done using hmmalign. You would need to create a FASTA file with all significant hits + your query. Chances are your query may already be in the list of significant hits. If not, this will do the trick using esl-reformat which comes with HMMer, and the output you already have from the -A option:

esl-reformat -u fasta output_msa > output_msa.fasta

Now open the output_msa.fasta file and add your query to it, followed by hmmalign:

hmmalign -o output_msa_new proteins.hmm output_msa.fasta

[el97004]

Thanks for this information! I was hoping it could be done within the hmmsearch command but this is a valid workaround. May I ask what you mean when you say:

Chances are your query may already be in the list of significant hits

Does hmmsearch actually include original query sequences into the output models?

[Mensur Dlakic]

Does hmmsearch actually include original query sequences into the output models?

I was assuming that my_protein_database is a modern protein database of good size. If so, it should have proteins that are at least similar if not identical to your query, even though they may come from a different organism and have a different accession number. hmmsearch can't output your query because your query is an HMM, so the only "query" it knows would be a consensus sequence emitted by that model. That consensus sequence is likely not identical to your query.

[el97004]

I see, makes sense! Thanks for the explanation!