Matching up homologous genes from different organisms with homologene

May 6, 2014

Homologene is an NCBI resource that constructs putative homology groups across species based on NCBI’s gene database. The resource can be queried interactively or through the eutils API. In addition, the data can be downloaded from the HomoloGene ftp site in a long form table with the following columns:

HID (HomoloGene group id)
Taxonomy ID
Gene ID
Gene Symbol
Protein gi
Protein accession

For example, here are the rows for the homologene group containing myc. Highlighted rows are H. sapiens (taxid 9696) and M. musculus (taxid 10090):

hid	taxid	geneid	gene symbol	protein id	protein accession
31092	9606	4609	MYC	71774083	NP_002458.2
31092	9598	464393	MYC	218563723	NP_001136266.1
31092	9544	694626	MYC	218847750	NP_001136345.1
31092	9615	403924	MYC	153070853	NP_001003246.2
31092	9913	511077	MYC	114050751	NP_001039539.1
31092	10090	17869	Myc	71834865	NP_034979.3
31092	10116	24577	Myc	71834866	NP_036735.2
31092	9031	420332	MYC	73661206	NP_001026123.1
31092	7955	30686	myca	153946419	NP_571487.2
31092	7955	393141	mycb	41055786	NP_956466.1

The following python script will create a table of all matches between human and mouse, including instances where there are 1-to-n mappings, rather than just 1-to-1 mappings

import sys
import collections
 
# taxids
MOUSE = 10090
HUMAN =  9606
 
def d():
    return {"h": [], "m": []}
table = collections.defaultdict(d)
 
for i, line in enumerate(open(sys.argv[1])):
    if i == 0:
        continue
    hid, taxid, geneid, symbol, _, _ = line.split("\t")
    taxid = int(taxid)
    if taxid == MOUSE:
        org = "m"
    elif taxid == HUMAN:
        org = "h"
    else:
        continue
    table[hid][org].append(geneid)
print >>sys.stderr, "read %i homologenes" % len(table)
 
print "hid\tmouse\thuman"
skipped = 0
for hid, hgene in table.items():
    # omit the human genes that don't have a mouse entry
    # and mouse genes that don't have a human entry
    if len(hgene["h"]) == 0 or len(hgene["m"]) == 0:
        skipped += 1
        continue
    for mouse in hgene["m"]:
        for human in hgene["h"]:
            print "%s\t%s\t%s" % (hid, mouse, human)
print >>sys.stderr, "skipped %i entries" % skipped

In the case of human to mouse, HomoloGene build 67 had 17461 mappings for 16784 homologous groups.