Hi I am wondering whether there are any good approaches for creating a census motif from a number of TFs in the same family, that are very closely related. For instance from Hox TFs or ZNFs, so if you have PWMs for individual members what is the best way to go about creating a consensus sequence? Thanks!
You can cluster the PWMs.
One option is to use GimmeMotifs (disclaimer: I wrote the software). The command gimme cluster can be used. There are also other tools that can do this, such as RSAT.
Do you want to make a sequence logo? You can do that easily through a web browser: http://weblogo.threeplusone.com/create.cgi
You can also do it with command line tools. Or are you trying to do something different?
I've also used PWMclus in "CCAT: Combinatorial Code Analysis Tool for transcriptional regulation." for a task like this with good results. I will warn that this type of analysis wouldn't work well for many families (especially Zinc Fingers) with very diverse motifs if you choose a single consensus motif. You have to choose 1 consensus per cluster.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
No actually its not just the logo, trying to figure out how to combine the PWM information from the different closely related motifs into one consensus motif.