You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
MODISCO produces 3 versions (representations) for each motif it learns
- PFM (position frequency matrix)
- CWM (contribution weight matrix)
- hCWM (hypothetical contribution weight matrix)
You can get a more intuitive explanation of these in the BPNet paper
https://www.nature.com/articles/s41588-021-00782-6
The PFM versions of each motif are classic motifs capturing frequencies of
bases at each position. So you can easily compare these to other motifs if
you like using standard motif comparison tools.
We prefer using the CWM representation as the primary motif representation
since it poses several advantages over PFMs. Again see the BPNet paper for
more details.
On Tue, Jun 21, 2022 at 12:24 PM sandraleebatista ***@***.***> wrote:
What are recommended ways to compare motifs from TFMoDisco and other
methods in terms of similarity and significance?
—
Reply to this email directly, view it on GitHub
<#103>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABDWEM6AGGFZPDPR4HPPCDVQIJGHANCNFSM5ZNPTPLQ>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
What are recommended ways to compare motifs from TFMoDisco and other methods in terms of similarity and significance?
The text was updated successfully, but these errors were encountered: