Transcription factor binding site prediction

2014-04-23T17:22:33+09:00

i am not a info guy but i feel relaxed when i read this

2014-05-21T19:44:34+09:00

Hi, Many thanks for this post. It seems difficult to screen for TF binding sites (particularly such as VDR). I am really new to bioinformatics and biostatistics (and still at the very beginning of the learning curve), but I cannot wait to give it a try. Very interesting blog.

Best,
Vassil

Reply

2014-05-22T00:06:53+09:00

Hi Vassil,

thanks! Here’s a classic paper on the identification of TFBSs -> Applied bioinformatics for the identification of regulatory elements. One of the points of the paper is the “futility theorem”, which is essentially all predicted TFBSs will have no function. This is mostly true for ChIP-seq peaks as well; it’s a noisy technology.

Good luck with the bioinformatics quest!

Cheers,

Dave

Reply

2014-07-02T21:16:12+09:00

Thanks for a useful review of options for mapping TFBSs.

I think the biggest trouble in this area of bioinformatics is having ready access to the necessary position weight matrices. That’s really what’s behind so many algorithms missing the Nanog binding site in your examples. Your solution (to generate your own PWM) is great if you know what TF you are looking for, but for discovery of putative TF binding sites in raw sequence data, we are all stuck finding whatever is available in the tool/algorithm we use, and missing anything else that might be present but not listed in that tool’s underlying data matrices.

I recognize that’s not your problem to solve; just musing on the possibilities and challenges in what you present here.

Reply

Pingback: MARCO transcription factor study | onlinenote

2014-08-15T22:08:17+09:00

Very nice post. I was oddly relieved that I’m not the only person struggling with what should be a simple task.

I was wondering if you knew about LASAGNA
http://www.biotechniques.com/BiotechniquesJournal/2013/March/LASAGNA-Search-an-integrated-web-tool-for-transcription-factor-binding-site-search-and-visualization/biotechniques-341032.html
It has a pretty friendly user interface and looks like you can use either established databases (e.g. JASPAR Core) or user defined PWM.

Reply

2014-11-22T04:04:14+09:00

Thanks, I was looking for a tool like this!

Reply

2014-12-31T08:41:01+09:00

Very nice post. And give me lots of useful advise. I’m sorry, but i didnt get if the probabilities of each base should be set as the test sequence or uniform background or control sets?

Reply

2015-01-05T05:59:18+09:00

For TESS, the default is to use a uniform frequency.

Reply

2015-01-07T08:27:23+09:00

But which is better, base compositions of test sequence or the default frequency?

Reply

	#!/bin/env perl

	use strict;
	use warnings;

	my $usage = "Usage: $0 <matrix.dat>\n";
	my $infile = shift or die $usage;

	my $accession = '';
	my $start = 0;
	my @entry = ();

	open(IN, '<', $infile) \|\| die "Could not open $infile: $!\n";
	while(<IN>){
	chomp;
	if (/^AC\s+(.*)$/){
	$accession = $1;
	}
	if (/^NA\s+(.*)$/){
	$accession .= "_$1";
	}
	if (/^P0/){
	$start = 1;
	$entry[0] = ">$accession";
	next;
	} elsif (/^XX/ && $start == 1){
	$start = 0;
	foreach my $line (@entry){
	print "$line\n";
	}
	print "<\n";
	@entry = ();
	}
	if ($start == 1){
	my ($junk, $a, $c, $g, $t, $junk2) = split(/\s+/);
	$entry[scalar(@entry)] = join("\t", $a, $c, $g, $t);
	}

	}
	close(IN);

	exit(0);

Transcription factor binding site prediction

TFSEARCH

PROMO

TFBIND

ConSite

Match

TESS

Using JASPAR

Conclusions

Like this:

Related

Leave a Reply Cancel reply