FIMO duplicated occurrences

已查看 85 次
跳至第一个未读帖子

Rocky Parida

未读,
2016年12月20日 12:35:402016/12/20
收件人 MEME Suite Q&A

Hi All
I am using FIMO to find all the significant+non-significant occurrences of a motif in a set of input sequences.

Sometimes I am getting an isolated motif occurrence and other times I am getting overlapping motif occurrences. Meaning motif #1 and motif #3 occur at the same position in the sequence and with same exact start and end.
Please comment. Is there a better way to resolve this?

My analysis:
I have tried looking at two metrics, motif occurrence significance q-value <0.05 and score.
Most of the time in these cases motif occurrence with highest score also has lowest q-value so I prefer that motif occurrence above all the other motifs that occur at the same spot of my sequence.
Sometimes nearly, 3% of the time motif with highest score disagrees with motif that has lowest q-value or vice-versa. This is where I am stuck.

New approach:
I am thinking to resolve this by preferring the motif occurrence that is most similar to the upstream sub-sequence, meaning the motif consensus that matches the most to the sub-sequence in question. 
 
Another scenario is motif #1 and motif #3 occurrences overlap with each other but not entirely, sometimes the overlap can be 1 base pair or 3 base pair or 4 base pair. It varies. My question is if this is a problem when I consider there are 2 motifs at this location with overlapping occurrences or is this wrong to say that.
Please comment.

Thanks in advance.
Rocky

CharlesEGrant

未读,
2016年12月20日 18:04:452016/12/20
收件人 MEME Suite Q&A
The FIMO raw scores for two different motifs are not directly comparable. You should use the p-value or the q-value to compare motif matches.

Depending on the exact nature of the analysis you are trying to do, there may be nothing to resolve. Biologically, a single site may bind to multiple transcription factors. In fact, the competition between TF for attachment to the binding site may be a key element of the regulatory system.

Rocky Parida

未读,
2016年12月20日 20:46:022016/12/20
收件人 MEME Suite Q&A
Thank Charles. Appreciate your answer and suggestion.
Rocky
回复全部
回复作者
转发
0 个新帖子