Question

How to cluster (nearest neighbour method) amino acid peptide sequences based on sequence identity?

0

Entering edit mode

9.3 years ago

vsivaranjanipriya ▴ 30

Hi,

I have a list of amino-acid peptides. For example:

ILM
FILM
FILVM
..
..

Now I need to cluster (nearest neighbour method) these peptides based on sequence identity

Please suggest any tool to do clustering

Thanks in advance

alignment • 4.7k views

ADD COMMENT • link updated 2.1 years ago by Ram 43k • written 9.3 years ago by vsivaranjanipriya ▴ 30

0

Entering edit mode

cd-hit, uclust

ADD REPLY • link updated 2.1 years ago by Ram 43k • written 9.3 years ago by Prakki Rama ★ 2.7k

0

Entering edit mode

Hi,

Try OthoMCL

http://www.orthomcl.org/common/downloads/software/v2.0/

ADD REPLY • link updated 2.1 years ago by Ram 43k • written 9.3 years ago by Juke34 8.5k

Ram · Answer 1 · 2015-01-26

If it is just the matter of sequence identity, uclust/cd-hit would probably be the best. If you want to take into account biochemical properties of amino acids used, there are some more sophisticated tools out there, namely: MUSI and the Gibbs sampling tool (IMHO better than MUSI) and also Hammock (there is also a Galaxy version of it), which is suitable especially for very large datasets.