I have a gene of of interest that I would like to compare between homologues. How does one go about finding a gene from known coding sequences across phyla? Afterwards I imagine I could do a Clustal sequence alignment to see how the sequences match.
5 Answers
There are various databases of homologs, for example:
- eggNOG
- Ensembl Compara (accessible via BioMart, I think)
- Homologene
- OMA
- TreeFam
The advantage of using an existing database is that more sophisticated methods for detecting orthologs than simple BLAST searches have been used (see "Computational methods for Gene Orthology inference" for a review), and everything is already precomputed, so it's much faster. The downside is that all precomputed methods need to use a snapshot of the genomes from the past, so not all currently available genomes will be there.
If I understand your question adequately, the Genome Browser at UCSC is a great place to start. If you know the name of the gene, you can search for it. For example, here is the page for human insulin receptor 1. From there you can compare 46 genomes, with alignments.
-
$\begingroup$ Wow that is an ugly interface. How do I use that thing? I think that's what I'm looking for. Looking at the INS-R1 gene as an example, I see chicken, zebrafish, a cluster of 46 vertebrates. Can I use this tool to create a consensus sequence of the conserved regions? $\endgroup$– user560May 13, 2012 at 2:53
If you don't have any idea where your gene will have a match with other genes, try something like Blast at the NCBI website.
This will give you a list of hits that you can then use to align with a multiple sequence aligner (MSA). The same NCBI page can give you a tree reconstructed from the results of the searching process, although there is a variety of methods that can be used if you just download the sequences and attempt to build the gene family alignment yourself using "[x] Select All -- Get selected sequences" in the NCBI blast results page, then downloading them in FASTA or other format with "Send to -- File -- Format FASTA -- Create File".
If what you want instead is to include your gene sequence into the best aligning place in an existing gene family alignment, you can try PAGAN.
maybe you'd like NCBI/Homologene instead. http://www.ncbi.nlm.nih.gov/homologene
or just using psiblast against NR if you have a nucleotide or protein sequence.
Just adding my 5 cents, there is a recent database called metaphors:
MetaPhOrs is a public repository of phylogeny-based orthology and paralogy predictions that were computed using resources available in seven popular homology prediction services (PhylomeDB, EnsemblCompara, EggNOG, OrthoMCL, COG, Fungal Orthogroups, and TreeFam). Currently above 306 millions of unique homologous protein pairs are deposited in MetaPhOrs database. These predictions were retrieved from 705 123 phylogenetic trees for 829 genomes. For each prediction, MetaPhOrs provides a Consistency Score and Evidence Level describing its goodness, together with number of trees and links to their source databases.
It is particularly good because:
It is the only homology database I know of that gives information on paralogy vs orthology.
It is linked to UniProt names
It provides a pretty good python API for batch queries