How Can I Use Tomtom (From Meme Motif Suite) With My Own Custom Database Of Motifs?

Question

[margiezilla]

Hi all,

It seems like it should be possible to use the TOMTOM software (locally installed and running via command line) to quantify motif similarity between motifs discovered via MEME/DREME and a custom set/database of motifs. I am wondering if anyone knows how to do this.

I can see that tomtom runs a command like this: tomtom -oc tomtom_example_output_files -min-overlap 5 -dist pearson -evalue -thresh 10 -no-ssc STRGGTCAN.meme JASPAR_CORE_2009.meme

in this example the STTRGGCTCAN(?) and JASPAR CORE 2009 databases being specified. Are these .meme databases consisting of motifs in the minimal MEME motif format? I have a custom set of thousands of PFMs in the minimal MEME format, can I run TOMTOM to compare my experimental motifs against these? I was trying to look at the source files of my MEME installation to look at where the .meme database files are being kept but I haven't been able to locate them.

Link to MEME suite: http://meme.nbcr.net/meme/documentation.html
TOMTOM paper: http://genomebiology.com/2007/8/2/r24
Minimal MEME format: http://meme.nbcr.net/meme/examples/sample-dna-motif.meme-io

Many thanks for any advice you can give :)

Answer 1

[margiezilla]

I figured it out! I thought the .meme files were something specialized, but STRGGTCAN.meme is literally just the minimal MEME format (http://web.mit.edu/meme_v4.9.0/doc/examples/tomtom_example_output_files/STRGGTCAN.meme) and the JASPAR_CORE_2009.meme is literally just another MEME file as well.

So you can just literally run:

tomtom -oc trial -min-overlap 5 -dist pearson -evalue -thresh 10 -no-ssc motiflist.meme databaselist.meme

and it generates a tomtom.html, tomtom.xml, and tomtom.txt

I was able to run my experimental motifs against a custom motif curation set.

Answer 2

[Neilfws]

It does sound as though installation of a custom database should be possible. There's a section in the manual titled Installing motif databases with some information about file paths.

There are also conversion utilities that convert to meme format, so it may be possible to go via an intermediate format, if you can get your data to look like one of those formats.

For more detailed advice I would contact the MEME developers; they are helpful and responsive in my experience.