parsing hammer 2.3.2 results using biopython

Question

[frcamacho]

Hi there,

I want to use SearchIO in Biopython to parse hmmer results. I have seen perl scripts, but I am not familiar with perl to understand how SearchIO works. Right now, what I am doing is a sorta hacky way, but running the hmmersearch and then outputing the results in a temp file to then parse the results. If anyone can explain how searchIO works in python and possible show some code that would be much appreciated.

Thanks!

Answer 1

[frcamacho]

Sorry for the late response, but this helped a lot!

To access the hmmer parsers from Biopython use: from Bio import SearchIO

with open (hmmFile,'rU') as handle: 
    for record in SearchIO.parse(handle, 'hmmer3-text'):
        query_id = record.id #seqID from fasta 
        hits = record.hits 
        num_hits = len(hits) # how many pfam hits per queries? 

        if num_hits > 0: # if there are more than 0 hits per query then we need to extract the info 
            for I in range(0,num_hits): 
                hmm_name = hits[i].id # hit name
                hmm_description = hits[i].description # hit decription 
                current_evalue = hits[i].evalue # evalue of hit

    handle.close() # close file

[frcamacho]

For more attributes of HMMER object to parse look at post: Biopython SearchIO HMMer 3 parsing, HMM model length?

[lwu8]

Hi, I am trying to using this script to parse hmmer results, but every time, I run it, it show for "hits=record.hits": AttributeError: 'Hit' object has no attribute 'hits' could you help me?

[frcamacho]

Hi, this script works with hmmer3-text format (https://lifebit.gitbook.io/cloudos/pipelines-documentations-and-examples-1/docker-pipelines-and-tools/hmmer) Is this the format you are trying to parse? If not, SearchIO has the following options, you can change the code to other forms

for record in SearchIO.parse(handle, 'hmmer3-text'):

hmmer3-domtab

1.61    1.61    1.61    HMMER3.0 domain table output format. The name hmmer3-domtab per se is in fact not used, since the program name has to be specified. For example, when parsing hmmscan output, the format name would be hmmscan-domtab.

hmmer3-tab  1.61    1.61    1.61    HMMER 3.0 table output format.

hmmer3-text 1.61    n/a 1.61    HMMER 3.0 plain text output format.

hmmer2-text 1.61    n/a 1.61    HMMER 2.x plain text output format.

[lwu8]

Thank you for answering my questions

Answer 2

[Peter]

There is an entire chapter on Biopython's SearchIO in the Biopython Tutorial, with examples, so I urge you to try that:

• http://biopython.org/DIST/docs/tutorial/Tutorial.html
• http://biopython.org/DIST/docs/tutorial/Tutorial.pdf

And yes, you can run hmmersearch, save the output to a file, parse the file. Or (and this is more advanced), you can try running hmmersearch from Python with the output to stdout, and parsing that directly within Python.