Finding arbitrary sequences in Seq objects from Biopython

Question

[knpayne2]

Let's say I have a large database of cdna sequences in the FASTA format, and I would like to identify a motif in the corresponding amino acid sequence. Let's say I need to find something like:

CxxCxxxxxxxxxxxxHxxx$

where $ will be H or C

I imagine one would start by parsing the fasta files, find the sites where these sub-sequences have to be, then

translate the corresponding coding DNA sequence, then I end up with an amino acid sequence that contains a sequence of this form. If I had a specific amino acid sequence in mind, I could easily find it by using the .find() method in the biopython module. However, I'm not sure how one can try to identify a form like above, in which one would search for a set of motifs.

Thanks!

[Ram]

The questions needs clarity at quite a few places. To start with, you have a database of sequences of which type? "FASTA" is the format, gives us nothing on the type of the underlying sequence.

[knpayne2]

Sorry, these are cdna sequences that are parsed from a set of FASTA files. Then I translate the sections between the KpnI and BamHI sites. With the amino acid sequence, I then need to find a sub-sequence that matches the pattern:

CxxCxxxxxxxxxxxxHxxx$

where $ will be H or C

I hope that is more clear.

Answer 1

[Asaf]

You can start with reading the sequences using: fain = SeqIO.parse('filename.fa', 'fasta'), then iterate the sequences: for seqrc in fain: and for each sequence translate() it and use re (regular expression) to find your pattern.

[cyril-cros]

BioPython now has a Bio.Motifs package, the MEME suite can help you scan for a list of motifs.

[Peter]

Note Bio.Motif was deprecated, you'd want Bio.motifs (lower case with an s).

[Peter]

Yes, I'd also have considering using a regular expression (via import re) here.