An integrative modular approach to systematically predict gene-phenotype associations

BMC Bioinformatics

Table 1 Highest confidence gene-phenotype association predictions stratified by phenotype.

Phenotype	Cutoff	Gene	Gene Description	PMID	Relevant GO Term
Urologic Diseases	1.00	ADRB3	adrenergic, beta-3-, receptor	18311486
Phagocytes	1.00	PML	promyelocytic leukemia		myeloid cell differentiation
Genital Neoplasms, Female	1.00	EXO1	exonuclease 1	15328369	DNA repair
Inflammation	1.00	CXCL2	chemokine (C-X-C motif) ligand 2	12892904	inflammatory response
skin disorder	0.99	SC4MOL	sterol-C4-methyl oxidase-like
Lymphocyte	0.99	FPRL1	formyl peptide receptor 2	15625007	G-protein coupled receptor protein signaling pathway
monocyte	0.99	TAP2	transporter 2, ATP-binding cassette, sub-family B (MDR/TAP)	12234057	immune response
Bone Marrow Cells	0.99	CXCL11	chemokine (C-X-C motif) ligand 11	15102366	immune response
Epithelium	0.98	ITGB4	integrin, beta 4	15194479	cell adhesion
Neoplasms	0.98	HMHA1	histocompatibility (minor) HA-1	14502255	intracellular signaling cascade

Each phenotype is listed with the predicted gene, the gene's description, and the proportion of decision trees from the random forest classifier that predicted its association to the phenotype. The final two columns provide supporting evidence in the form of a PubMed ID for a published work or a Gene Ontology annotation that is consistent with the phenotype.

ISSN: 1471-2105