Compare sequence with MSA consensus

Question

[rickbeeloo13]

I have done a multiple sequence alignment for two proteins with similair sequence, however there are some minor differences which discriminate a cytoplasmic form from a mitochondria form. I downloaded the output of the MSA alignment as a consensus sequence and now I want to "blast" this against a genome to find which of these two is probably the version present in the genome.
Which method/tool can be used to compare a sequence with a set of other protein sequences (MSA)?
it's not an option to blast just one cytoplasmic version and one mitochondria version and see which of these gives the best matches, an MSA is/was neccesarry

Answer 1

[ddiez]

Basically you want to construct a model or profile of your two types of proteins and use it to search similar proteins. Using the consensus sequence is not very good for this task.

One option is to use HMMER. Basically, you use your MSA to construct a HMM profile. This profile will capture the positions that are similar. Then, you can use this profile to search for proteins matching the profile in sequence databases. For example, if you have the protein sequences of a target species you can search with the two profiles (one for the mitochondrial and the other for the cytoplasmic). BLAST can perform a similar task with the PSI-BLAST tool.

[rickbeeloo13]

Thankyou! I have one problem using HMMER when I use this command:

hmmbuild --amino --informat "CLUSTAL" "hmmCytopolasm.hmm" "input.clustal"

I get the error:

Error: CLUSTAL is not a recognized input sequence file format

Any idea what I did wrong here?
I can convert it to Stockholm using this converter however I'm not sure how reliable this is

[ddiez]

What version of HMMER are you using? Might be that you need to specify "Clustal" in lower case? I would try not including that parameter (default is to autodetect).

[ddiez]

I just tried with a small test file and HMMER (3.1b2) guesses correctly the format without specifying it. Also didn't need to use --amino as that is guessed from the input format too. And, it also works with --informat "CLUSTAL". Your problem is that there is a whitespace in CLUSTAL, i.e. you have --informat "CLUSTAL ".