7N Motif Search Over The Genome

Question

[PoGibas]

I do have really short word size (microRNA target sequence).

Want to search enrichment of those motifs in my DNA seq & lots of randomly simulated same length genome sequences. (I am going to RNA->DNA before the search).

What way, what tool should I use for such short motif search?

I have heard about Vmatch, but maybe there is a free software?

Really looking forward to your answers and suggestions...

PS.: Or any simple pl script (within y.fa search x motif) would work fine.

Answer 1

[Farhat]

You can use the following script for that. The usage is perl patt_search.pl fasta_file.fa AATTATA TATA ... if you save the script as patt_search.pl. You can give any number of motif sequences. It will recognize IUPAC DNA ambiguity codes. The output is a bit weird because I used it as a feed into another program but it looks like this.

{"chrX:6362554-6365728",{{"TAATTA"}, {260, 2466, 2875}}, {{"CCCCCCCC"}, {1412}}},
{"chrX:6379561-6405165",{{"TAATTA"}, {275, 776, 1048, 1226, 1722, 2753, 3585, 3644, 4951, 5084, 11164, 12712, 16259, 17695, 18211, 18574, 18745, 19204, 19838, 19859, 21405, 23529, 23740, 24372}}, {{"CCCCCCCC"}, {4536, 5673, 9148, 12449, 14132, 16375, 20132, 20140, 21463, 21471, 21975}}},

It contains the fasta header followed by the motif searched for followed by all the locations that it was found on within that sequence. The program can be downloaded from https://github.com/Farhat/patt_search

ETA: Now it can handle more complicated DNA character strings like TTA{3,7}T and their corresponding reverse complements.

[PoGibas]

Thanks! You saved me two days at least! :) Does it do rev/comp too?

[Farhat]

Yes, it will search for reverse complements too. You can also use IUPAC ambiguity codes and N to match any base.

[k.nirmalraman]

Does this code also find patterns like ACA{0,7}TG and detect patterns as follows in input stream ACAAAAAAATG, ACAATG, ACAAAATG be detected? and Does N for {A or T or G or C} also work?

As an extension I would like to ask if it is possible to read muliFasta file with the given header? It will be of great help, I can get that done!!

PS: I am not a perl person yet ;) would love to use the code just as it is and format the output to my need (basically a bed file), if it works!!

[Farhat]

No, it will not work for general regular expressions. The expansion for N isn't supported but it is a minor change. I'll edit the program to include that.

[k.nirmalraman]

Thank you very much!!

Just for the record, dna pattern match with some advanced option is available here as part of RSAT tool. However, one cannot integrate this to a analysis pipeline. I would like that... :)

[k.nirmalraman]

I was actually hoping I can extend this script a bit, to find character repetitions like I mentioned above i.e., ACA\{0,7\}TG to find ACAAAAAAATG and ACAATG and so on....

I added $patt =~ s/\d+/$&/g; to replace_ambiguous subroutine before the return statement.

Changed a bit of reverse complement to $revcomp =~ tr/ACGTacgt[]{}N/TGCAtgca][}{./; to accomodate paranthesis { }.

What I end up searching in the FASTA file for reverse strand is a problem.

Eg., Input in argument : CR\{7,10\}N\{5,8\}ATGC

Generated Forward Strand Look Up: C[AG]{7,10}[ACGT]{5,8}ATGC

Generated Reverse strand: GCAT{8,5}[ACGT]{01,7}[CT]G

The reverse complement string is a problem.... I don't think there is a easy way to do it from my limited knowledge... May be can you help me to achieve this???

[Farhat]

This is indeed a bit more complicated but can be solved with regular expressions. You can download the modified program at https://github.com/Farhat/patt_search You will have to enclose your patterns in quotes when using it on the command line to prevent shell from parsing braces.