Multiple alignment filtering

Question

[guillaume.rbt]

Hi all,

I'm doing a multiple sequence alignment from protein sequences, and I would like to get only the varying positions, so that I can create sequence logo of variyng amino acids.

Would anybody a tool for multiple alignement that can output only variyng amino acids? Or maybe a tool that can filter a multiple alignment format, as clustal, to filter out all 100% identical amino acids between sequences?

Thanks

[GenoMax]

You can upload multiple sequence alignments to WebLogo to create a logo. If you are only looking for variable positions then it is opposite of what people are normally looking to do. Unless someone knows of an existing tool you may need to write something custom yourself.

[guillaume.rbt]

Thanks! The problem I've got is that I have long sequences (around 200 AA), with few variants (around 5). Hence if i submit the whole sequences to Weblogo the output it won't be readable, that's why I try to filter the multiple sequences alignment input.

[Joe]

If you filter out only the variable positions, you’ll lose the positional information though, or are you not worried where the variants are?

[guillaume.rbt]

Indeed I will have to keep track of the positions for my figure.

[Joe]

The scale will remain the same in that case. A 200 character web logo won't be readable, as you say, but there's no filtering you can do that will reduce the x axis if you're trying to keep the positions.