Regular Expression To Match A Codon Pattern

Question

[hadasa]

I have an amino acid pattern DIGD i would like to capture at the rna/DNA level. Each of thes Amino Acids are coded for by more than one codon. Thus to capture the pattern at the DNA/RNA level you need to match either of GAU,GAC followed by either AUU or AUC or AUA, followed by either GGU, GGC,GGA,GGG and so on.

I have a quick regular expressioon, which given the mrna should capture these patterns.

(gau|gac)?(auu|auc|aua)?(ggu|ggc|gga|ggg)?(gau|gac)

but just does not feel right! am i missing something. Someone with an improved approach?

Here is a second attempt

((:?gau|gac)(:?auu|auc|aua)(:?ggu|ggc|gga|ggg)(:?gau|gac))

Example strings to match string1

gauauaggagauaucguuagaggaaaagaucuauuuuaugguaauacacaugaaaguaa

gaauauauaugaaggauugucgaacaaugguguaaaagcucgcuacgaaggugauacug

gcgccacaguauggaaggcuaucacauguaaagcuaaggaagcugauaaauauuuuaga

gcgaagauacugcggcacauaaucgaaauagguugggggaucgguauuuggauug

The last string should not match

[Giovanni M Dall'Olio]

try to put an example of which strings you would like to match, and which you would like to avoid to match.

[Pierre Lindenbaum]

I tried GA[uc]AU[UCA]GG[UCAG]G(AC|GU) but it doesn't even match your first example. Are you sure of your pattern?

[Simon Cockell]

Your pattern isn't in any of your examples.

[hadasa]

Sorry pierre, to make it simple lets assume everything is in lowercase. Am more intrested in the knowing the correct regular expression than perfoming the match.

[Giovanni M Dall'Olio]

your example is yet not complete... you should put an example of the output you expect after the regex match. And why the latter example is wrong?

[Pierre Lindenbaum]

@biorelated I tested my pattern with the '-i' option of egrep, the upper/lower case have no importance here.

[Istvan Albert]

Note that regular expressions may have exponentially(!) degrading performance on certain inputs (unless using an engine designed to avoid it) - therefore you might want to investigate just translating to protein sequence and matching that. I learned this the hard way, here is a writeup.

[hadasa]

Yeah. The interest is actually on the coding sequence more than the protein. So i could have matched the translated sequence. so wanted to capture the actual codons rather than the resulting peptide. Thanks for the link

Answer 1

[Simon Cockell]

Oh wait, I get it, your first and second attempts in the question don't match one another. Codon 1 and codon 4 are the same?

EDIT: You really don't need to include the whole codon in your variable match, as only one or two bases vary. The following pattern works just fine.

import re

#p = re.compile('ga(:?u|c)au(:?u|c|a)gg(:?u|c|a|g)ga(:?u|c)')
#changed for readibility
p = re.compile('ga[uc]au[uca]gg[ucag]ga[uc]')
strings = strings = ['gauauaggagauaucguuagaggaaaagaucuauuuuaugguaauacacaugaaaguaa',
 'gaauauauaugaaggauugucgaacaaugguguaaaagcucgcuacgaaggugauacug',
 'gcgccacaguauggaaggcuaucacauguaaagcuaaggaagcugauaaauauuuuaga',
 'gcgaagauacugcggcacauaaucgaaauagguugggggaucgguauuuggauug',
 'gauauuggugau']

for string in strings:
    print(p.search(string))

Produces

<_sre.SRE_Match object at 0x24fde0>
None
None
None
<_sre.SRE_Match object at 0x24fde0>

As Giovanni says, checks for case and sequence validity would be a good addition, but are probably out of the scope of your question.

[hadasa]

Thanks a lot. Any tools you may know of that quickly generates or ca be used to validate regexps? :)

Answer 2

[Giovanni M Dall'Olio]

First you have to write some examples of sequences that should match your regex, and also some sequences that should not:

>>> good_ones = ('gauauugguggu', 
                  'gacauuggcggu', 
                  'gacauagggggu',
             )
>>> bad_ones = ('gaugauauugguggu',              # repeated codon
                  'auuggcgcu',                  # this case misses one codon
                  'gagauagggggu',               # wrong sequence
                  'acauuggcggu',                # this case misses one nucleotide
             )
>>> difficult_ones = ('GAUAUUgguggu',           # lower/upper case matters?
                  'gacattggcggt',               # T == U?
                  ' gacauagggggu ',             # leading spaces?
                  'ACGAGCgacauaggggguAGCTCGATCG',  # what if the sequence is within another sequence?
             )

Are these example correct? you should decide what you want to do in the latter examples: do you want to distinguish between lower and upper cases? between Us and Ts? what if the sequence has a leading or ending space?

Then, the regex you wrote is almost correct, but you can use a non-grouping operator (?: since you are not interested in making groups. Moreover, you should

>>> import re
>>> m = re.compile("^(?:gau|gac)(?:auu|auc|aua)(?:ggu|ggc|gga|ggg)(?:ggu|gac)$")
>>> for seq in good_ones: m.search(seq)
<_sre.SRE_Match object at 0x1666f38>
<_sre.SRE_Match object at 0x16665e0>
<_sre.SRE_Match object at 0x1666e68>

>>> for seq in bad_ones: m.search(seq)

>>> for seq in difficult_ones: m.search(seq)
???

[hadasa]

within the last set of codons, the firt codon should be gau|gac not ggu|gac :) thanks

[hadasa]

more intrested in the correctness of that regular expression. Since i wonna check out the indexes where the matches starts.. :) thanks!