Re: appropriate E-value threshold for filtering the MEME result

689 views
Skip to first unread message

James Johnson

unread,
May 5, 2013, 10:33:41 PM5/5/13
to meme-...@googlegroups.com
MEME will keep finding motifs until it reaches the motif count you specify. In this case the motif is almost certainly not significant and it should be ignored. I would not consider anything above a E-value of 0.05 significant.

seu...@gmail.com

unread,
May 8, 2013, 3:03:00 AM5/8/13
to meme-...@googlegroups.com
Thanks for reply.
I am wondering is there any article or material to which was referred you use 0.05 as threshold.

James Johnson

unread,
May 8, 2013, 3:20:11 AM5/8/13
to meme-...@googlegroups.com
If you're wondering why a p-value or E-value of 0.05 is the de-facto standard cut-off then I suggest you read the history section of the p-value article on Wikipedia:
https://en.wikipedia.org/wiki/E-value#History

If you really want a citation then just do a literature search for someone else who has done motif discovery and read their paper to find out what they used.

Hima vyshnavi.a.

unread,
Jun 5, 2014, 12:23:03 AM6/5/14
to meme-...@googlegroups.com
HI,
    i would like to know which parameter changes the result?? E-value? frequence or the scores in the matrix? If i want to develop another tool which parameter should i change?

CharlesEGrant

unread,
Jun 11, 2014, 7:08:37 PM6/11/14
to meme-...@googlegroups.com
Hi Hima,

I'm afraid you'll have to be more specific about how you want to change your results. MEME supports numerous command line options, all of which can change the results.
The command line options are described here:






Reply all
Reply to author
Forward
0 new messages