I am currently trying to find homologues of a protein I am working with, but BLAST has been giving me nothing useable. I have now found a dataset of 1500 protein sequences of potential candidates that I want to align to my reference sequence. I have tried Clustal, Mega, Muscle, MAFFT and pretty much everything under the sun, but with this many sequences and only limited experience, I am having trouble achieving what I want to do, as the programs simply crash or lock up after a few minutes..
Instead of the traditional multiple sequence alignment, where every sequence gets aligned to every other sequence with multiple iterations, I want all of the sequences from the dataset to only be aligned to my one reference sequence. Think of it as doing 1500 pairwise alignments only. What would be the best way to perform this kind of alignment?
You still be using MSA, but the perspective would be different. You will be having 1500 sequences + reference sequence in your query file. Better if you have first sequence as your reference (would make it distinctive). Then do a MSA without reordeing. And interpret the alignment file pairwise. In this way you will be able to see the alignment of your query against your reference.
Or
use the --add flag in mafft to add sequences to an existing alignment.
( https://mafft.cbrc.jp/alignment/server/add_sequences.html )
Or
use blast and get out the aligned sequence. Blast is highly customizable and you can tweek the output according to your needs.
Do pwa in jalview and retrieve the first n-1 alignments in the result file.
I think it is easy as you work with GUI.
MAFFT or hhblits can be used for such large set of data : http://toolkit.tuebingen.mpg.de/hhpred. you may want also to try bioperl: http://doc.bioperl.org/releases/bioperl-1.4/Bio/Align/AlignI.html
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology