Hi everyone,
I am currently trying to align several thousand (5-10) protein sequences to a related protein sequence (reference) to determine the percent identity of the large subset of sequences to the reference.
Does anyone know of a program that I can use to align large sets of protein sequences to a reference sequence. I know that Geneious is capable of performing reference alignments, but it is almost prohibitively expensive and I am capable of using command line tools.
If anyone knows of any open-source tools that could help me perform these tasks I would greatly appreciate the suggestions.
Thank you in advance!
This is old but not yet correctly answered. I am surprised how one can answer without even reading the question.
He asked about REFERENCE ALIGNMENT. Multiple sequence alignment is exactly NOT referenced.
AVID is capable of doing reference alignment.
Bray, N., Dubchak, I. and Pachter, L. (2003) AVID: A Global Alignment Program. Genome Res., 13:97
Please note that VISTA suite implements AVID as an option as alignment tool.
Best.
Hello Alex, Please use the multiple sequence alignment search tool called Tcoffee, an online resource which is open and freely accessible at the European Bioinformatics Institute Website. Please follow this link to access the Tcoffee software Online. http://www.ebi.ac.uk/Tools/msa/tcoffee/ Or you may want to download another progam called Clustalx or ClawstalW2 available at the EBI website: www.ebi.ac.uk/Tools/msa/clustalw2/ . Please let me know if this was helpful. Its my pleasure.
''determine the percent identity of the large subset of sequences to the reference''
Blastp may do some help...and just find a perl script from google or write it to acheive that
Hi Alex
Despite of the already pinpointed suggestions, I personally like MAFFT as aligner. It can be downloaded and used as standalone software. In a Linux box works like a rocket, and I think that it can be used to align up to 10,000 seqs.
Here is the link, and hope it helps. Best of luck. Paco
http://mafft.cbrc.jp/alignment/software/
For such a large set of sequence, I would use either MAFFT or hhblits (lightning fast ...)., use them online -> http://toolkit.tuebingen.mpg.de/hhpred.
Their results page provides a nice graphical plot which will answer your needs (sequence identity, amino acids conservation per position, etc).
If you want a more complete descript, you could use bioperl:
http://doc.bioperl.org/releases/bioperl-1.4/Bio/Align/AlignI.html
Another vote for MAFFT. There's a paper comparing the accuracy and performance of all multiple sequence aligners. MAFFT was arguably the best multiple sequence aligner covered in the paper.
In addition, it appears to provide a polished interface consistent with most common UNIX command line applications. Easy to install, easy to use, and fast.
The newest version of the clustal alignment programs, clustal Omega, is fast and accurate for large alignments. Their paper performs several benchmarks in which clustal Omega outperforms other tools suggested here. It is my tool of choice for alignments large and small, and is easy to install/use on a linux system.
http://www.clustal.org/omega/
http://msb.embopress.org/content/7/1/539
I tried MAFFT, but the program aligns one-to-one sequences. How to align sequences to only one reference sequence?
There are many cases where generic multiple sequence alignment programs go off on tangents when large numbers of sequences are fed into them. As the first post indicated, sometimes we just want to see everything aligned against a given reference sequence.
A few years ago I had written a program which converts a series of blasr -m5 format records (which may be many sequences aligned to one reference) to a fasta alignment. Blasr could only be used for DNA though. So a modified version of the SSWlib test program was written which would work for proteins and emit alignments in blasr's -m 5 format. It is SW alignment (so parts may not align), but it is very fast. Unfortunately it has a size limit of around 5000 residues, and it truncates to that without warning. The EMBOSS water or needle programs can be used instead - they are slower. Here is an example using needle:
#ref.fasta == protein sequence to which all others should be aligned
#many.fasta == many protein sequences. It should have a copy of the reference.
# NSEQ is the number of sequences in many.fasta
cat >needle_to_m5.sh
- Badges
- Science topic
- Similar topics
- References