AME vs MEME questions

[Kristina Sakers]

Hello,

I am using the AME webtool. I was wondering if there is anyway to see how many of the sequences have the motif that was found (similar to how MEME does this). Further, I see motifs when I scan with AME but these motifs are not found when I use MEME. Is there an explanation for this?

Thank you.

[CharlesEGrant]

AME and MEME are very different tools. 

MEME is performing de novo motif discovery. It is considering only the sequence data you provide, and looking for short, similar, sequences that occur more often than would be expected by chance. The short, similar sequences are used to construct a position-specific weight matrix (PWM) that characterizes the motif.  MEME reports the PWM, and also reports the best matches to that PWM in your input sequences

AME is testing your sequence data for motif enrichment, that is, do your sequences contain more or better matches to the motifs then would be expected by chance. You provide the motifs to be matched as input. A few matches to a motif may be enough to establish enrichment, but not enough to support de novo 'discovery'.

is anyway to see how many of the sequences have the motif that was found

In general you can use FIMO (Find Individual Motif Occurrences) to obtain a table of motif matches in a set of sequences. 

However, depending on which AME scoring method you used, this question may not be valid. The default motif scoring method for AME is to evaluate the matches to a motif at every position in the sequence and then take the average match score. 

If you selected the "Total matches" scoring method in the "Advanced Options" settings, AME will score sequences by counting the number of motif matches that pass the provided p-value threshold. You could set the same p-value threshold in FIMO, and use it to generate a list of matches.