Table 2.

AUROC with genomic features and clinical severity

Dataset	Score type	Severity alone	Duke		Sage LR		Sage RF		Stanford
			Gene model alone	Joint model	Gene model alone	Joint model	Gene model alone	Joint model	Gene model alone	Joint model
Discovery datasets
EMEXP3850	PELOD	1	0.947	1	0.916	1	1	1	1	1
EMTAB1548	SOFA	0.735	0.817	0.843	0.863	0.87	1	1	0.849	0.863
GSE10474	APACHE II	0.551	0.53	0.626	0.682	0.758	1	1	0.722	0.697
GSE27131	SAPS II	1	0.7	1	0.7	1	1	1	1	1
GSE32707	APACHE II	0.546	0.514	0.537	0.712	0.702	0.996	0.996	0.81	0.805
GSE63042	APACHE II	0.774	0.679	0.797	0.866	0.868	1	1	0.742	0.815
GSE66099	PRISM	0.781	0.806	0.84	0.916	0.913	1	1	0.881	0.892
GSE66890	APACHE II	0.723	0.802	0.847	0.711	0.759	1	1	0.834	0.849
Validation datasets
EMTAB4421	APACHE	0.705	0.695	0.771	0.81	0.762	0.714	0.752	0.829	0.838
GSE21802	SOFA	0.812	0.333	0.833	0.708	0.792	0.583	0.833	0.75	0.833
GSE54514	APACHE	0.776	0.936	0.944	0.701	0.739	0.902	0.927	0.816	0.825
HAI datasets
Glue Burns D1-D30	Denver score	0.482	0.808	0.842	0.721	0.731	0.606	0.604	0.74	0.756
Glue Trauma D1-D30	MODS score	0.927	1	1	0.938	0.979	0.667	0.958	1	1
UF P50 12H	SOFA	0.941	0.573	0.945	0.652	0.945	0.6	0.952	0.682	0.945

Some gene model AUCs may differ from Supplementary Table 2 since samples without severity scores were dropped from this analysis