This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Python'S Regular Expressions To Find All Instances Of A Codon In A Sequence

I am trying to find all instances of a specific codon in a given gene or RNA sequence using Python's regular expressions. The findall function seems to be able to do the the job, however, the problem is that one needs to match codons not any subsequent three letters, which may not be part of a codon. Here's an example:

>>> seq='CTCTTACTT'
>>> import re
>>> re.findall(r'CTT',seq)
['CTT', 'CTT']

The first CTT that it finds does not correspond to a codon (CT**CTT**ACTT) since we have only three codons in the given sequence including: CTC, TTA,CTT

Obviously, the most straightforward way is to use a loop, extract each codons from the sequence and compare it with CTT (the codon we are searching for), but I am looking for a smarter way of doing so.

programming python sequence
• 9,671 views • link
updated by Biostar Biostar • written by nobodyknowsme57 • nobodyknowsme57

5 answers

Try [m.start() for m in re.finditer('(?=TTT)', 'CTTTTTGTA')]. It will return overlapping instances. In this case: [1,2,3]

• 0 views • link
written by Ashutosh Pandey Ashutosh Pandey

There will certainly be a more "clever" way of doing it, but for mine, this approach with a generator function and list comprehension gets the job done while being readable:

def codons(seq, frame):
     """Generator function that yields DNA in one-codon blocks 

     returns a tuple containing (codon, position relative to start)
     note: reading frame is 1-based, index for the nucleotide position is 0-based
     """
     start = frame -1
     while start + 3 <= len(seq):
         yield(seq[start:start+3], start)
         start += 3

test = 'CTCTTACTT'
CTT_positions = [p for (c, p) in codons(test,1) if c == 'CTT']

Which will return [6]

• 0 views • link
written by David W David W

With re in python, you have mention specifically each pattern. Sometimes you might need to find a pattern with indels or snps. In that case Mummer and http://blog.theseed.org/servers/2010/07/scan-for-matches.html will be useful.

• 0 views • link
written by Pappu Pappu

[m.start() for m in re.finditer('CTT', 'CTCTTACTT')] will return all the indices. In this case it will return [2,6]. Now you can pick the second occurence as it follows the reading frame. Get all the indices that are fully divisible by three (also include 0).

Thanks

• 0 views • link
written by Ashutosh Pandey Ashutosh Pandey

Using finditer is a good idea but it's not gonna work all the times as it finds non-overlapping matches. Here's an example showing why this is not working:

[m.start() for m in re.finditer('TTT', 'CTTTTTGTA')] [1]

The actual position for the codon of interest here is is 3 and not 1, because the sequence is composed of the following codon: CTT TTT GTA

• 0 views • link
written by nobodyknowsme57 • nobodyknowsme57

Something like this should really appear as a comment under the answer it refers to - the "Answers" section is intended for, well, answers to the original question.

• 0 views • link
written by David W David W

Log in to answer this question.

More posts like this