MEME for large datatset

[Ashwini M]

Hi,
I am using MEME without any problem for input dataset of about 6 million bases. Whereas I am facing problem in datatset of about 17 million. The program halts mid-way without giving any error message.
In both cases the -maxsize is set to 150000000 and -motifs=1.
If it is the problem with data size should i use an alternative to MEME or is it ok if I split the input fasta file into 3 or 4 fasta files and run them individually.

It would be really great if anybody could help me with this

[rhinoceros]

Not that I've ever used this program, but did you try running it in verbose mode (quick check on documentation states that you do this by "-V")?

[Ashwini M]

Thank you for your response.
The -V option was not helpful.

[sphil]

Hi,
maybe I am late. But i think splitting the file is not an option. MEME does EM so you will change background frequencies and therefore the EM-Value itself. The only way this won't happen is when the letter frequencies [A,C,G,T] are the same within all generated files which hardly will be the case.

I would look for an alternative.

Hope that helps,

best

phil

[Ashwini M]

Hi,
Thanks a lot for your reply.
You are right. Splitting the file is not an option.

I was trying with meme-chip which can handle large data.
meme-chip by default takes random 600 sequences of 100bp each. But this is not what is required.

So, I set -ccut to 0 and -nmeme to 6623. But meme-chip throws an error "Dataset too large (> 100000). Rerun with larger -maxsize."
meme-chip does not have an option to set -maxsize

No. of sequences - 6623
Maximum length - 9925
minimum length - 367

[sphil]

hey,

try to check: http://www.biostars.org/p/57125/ or http://nar.oxfordjournals.org/conten...ar.gkr574.full.

best

phil