I first assigned M. tuberculosis genes to E. coli using KEGG orthology, then used BLASTP against the E. coli proteome to recover sequence-based homologues for genes missing KEGG annotations, producing a combined KEGG+BLAST homology table for downstream analysis. Is that normal that from 4000 genes, I only could map 1700 genes?
non human homology detection
1 answer
You will need to distinguish between orthologs (speciation) and paralogs (gene duplication).
With UniProt ID"s you should be able to search UniRef database to identify clustered members. e.g. https://www.uniprot.org/uniref?query=Q04657
NCBI's COG database would be another good source: https://www.ncbi.nlm.nih.gov/research/cog/
OrthoDB would be another source: https://www.orthodb.org/?
Failing that you will need to go the classic route of BLAST --> MSA etc.
That number should be higher I would think. I have not used this tool from KBase but you could give it a try: https://kbase.us/applist/apps/GenomeProteomeComparison/compare_two_proteomes/release
Log in to answer this question.
More posts like this
-
Ensembl ID conversion - other than Uniprot
written by NS •
-
GO Terms retrieval
written by brianbtarimo •
-
GO terms retrieval
written by brianbtarimo •
-
How to convert references ID list in UCSC reference list
written by Lila M
-
Retrieve gene name and sequence by Transcript ID
written by tianlu •
-
Pathview IDs conversion
written by friasoler •
-
mapping a list of affy IDs from citrus to arabidopsis locus IDs
written by zizigolu
-
From Ensembl protein ID to sequence
written by Joseph Hughes
-
Convert kegg ids to uniprot ids
written by Anusha •
-
[Cytoscape] Merge nodes by column/attribute
written by enricoferrero
