Data mining and machine learning approaches for the integration of genome-wide association and methylation data: methodology and main conclusions from GAW20

Table 1 Investigated data in contributions from the Data Mining and Machine Learning group at the GAW20

Contribution	Sample size	Real data	Simulated data	GWAS	EWAS	Investigated phenotype(s)
Random forest (Darst)	680		X	X	X	log average post-TG − log average pre-TG
Deep learning (Islam)	993/499^a	X			X	pre-TG, post-TG
Cluster analysis (Kapusta)	446	X		X	X	relative TG difference, metabolic syndrome
Mixed models (Datta)	680		X	X	X	post TG-pre TG
Gene-set enrichment (Piette)	680		X	X	X	log average post TG/log average pre TG

EWAS epigenome-wide association study, GWAS genome-wide association study, TG triglyceride concentration
^aThere were 993 participants in the Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study, but posttreatment methylation data was only available for 499

ISSN: 2730-6844