How to identify mismatch and gap positions

[pony2001mx]

Dear all,

I need to identify the homologs of many short sequences, and then make statistics of the mismatch and gap positions between each short sequence and its homolog. The first step can be readily done by BLAST, FASTA or patscan, but the second step is a little bit troublesome for me. I am wondering if it's feasible to parse the BLAST or FASTA result, and transform the alingment to mismatch and gap information using a perl script. Would you please suggest me some tools and methods to achieve this analysis? Is there a perl script to extract the information i need from BLAST/FASTA output?

Thank you very much!

[maubp]

Have you looked at BioPerl for parsing the output inside Perl?

[lindenb]

cross posted on http://www.biostars.org/p/89151/

[pony2001mx]

Thank you very much. I have sorted it out.

[GenoMax]

Thank you very much. I have sorted it out.

For the sake of completeness: a solution was provided on the biostars thread that was posted above.

pony2001mx: If you found an alternate (better) solution then please include it here.

[pony2001mx]

Hi GenoMax,

Lindenb's solution (https://github.com/lindenb/jvarkit/wiki/BlastnToSnp) should be a nice one. I am performing blast for short sequences and testing his java program.

I am a perl beginner. I believe parsing blast or fasta output using perl should be another solution. There are handful of websites including SEQAnswers that describe methods parsing blast output, for examples,

http://seqanswers.com/forums/showthread.php?t=6484

http://www.bioperl.org/wiki/Parsing_BLAST_HSPs

http://www.1cro.com/Bioinformatics/129.htm

Actually, my case may be a little bit more complicated. I used Patscan to identify miRNA targets, the output is like below.

miRNA_ID    miRNA_sequence                            target_sequence
miR1            AUGCGGUGUACACGUACGUCG        AGGCGGUGUACACGUACGUG
......................

miR1 and its target has one mismatch (the second nt) and one insertion (the 20th nt). I need to make statistics of this information, which has not been given by Patscan.

I plan to perform blast using miRNA sequence as query and its target as subject, and then use Linenb's java program to identify SNPs.

Honestly, I am testing and my problem has not been completely sorted out. But I think i am closer. If you and others have better solutions or ideas, please feel free to let me know. I believe it is a general question for many biologists. Really thank you very much!