An investigation of biomarkers derived from legacy microarray data for their utility in the RNA-seq era

Su, Zhenqiang; Fang, Hong; Hong, Huixiao; Shi, Leming; Zhang, Wenqian; Zhang, Wenwei; Zhang, Yanyan; Dong, Zirui; Lancashire, Lee J; Bessarabova, Marina; Yang, Xi; Ning, Baitang; Gong, Binsheng; Meehan, Joe; Xu, Joshua; Ge, Weigong; Perkins, Roger; Fischer, Matthias; Tong, Weida

doi:10.1186/s13059-014-0523-y

Table 7 Performance metrics for the assessment of cross-platform transferability of signature genes of predictive models based on the TCGA AML data

From: An investigation of biomarkers derived from legacy microarray data for their utility in the RNA-seq era

Algorithm	Direction	Endpoint	Gene set	Microarray model				RNA-Seqmodel				T-index
				Accuracy		AUC		Accuracy		AUC
				Mean	95% CI	Mean	95% CI	Mean	95% CI	Mean	95% CI
k-NN	Forward	Sex	A	0.988	0.988-0.988	0.987	0.987-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.988
			B	0.980	0.953-0.988	0.980	0.955-0.987	0.986	0.977-0.988	0.985	0.977-0.987	0.986
			C	0.496	0.419-0.57	0.496	0.413-0.579	0.734	0.419-0.988	0.733	0.418-0.987	0.647
		Cyto Risk	A	0.945	0.907-0.988	0.895	0.805-0.993	0.947	0.872-0.988	0.905	0.794-0.993	0.947
			B	0.939	0.872-0.977	0.904	0.745-0.986	0.942	0.884-0.977	0.925	0.825-0.986	0.942
			C	0.919	0.848-0.965	0.884	0.767-0.964	0.907	0.872-0.953	0.854	0.776-0.948	0.908
	Reverse	Sex	A	0.988	0.988-0.988	0.987	0.987-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.988
			B	0.508	0.419-0.599	0.501	0.41-0.591	0.506	0.407-0.593	0.508	0.409-0.593	0.507
			C	0.977	0.977-0.977	0.977	0.977-0.977	0.988	0.988-0.988	0.987	0.987-0.987	0.977
		Cyto Risk	A	0.937	0.837-0.988	0.886	0.724-0.993	0.951	0.843-0.988	0.924	0.815-0.993	0.938
			B	0.856	0.744-0.919	0.758	0.59-0.888	0.848	0.767-0.907	0.748	0.589-0.893	0.855
			C	0.930	0.86-0.965	0.887	0.767-0.979	0.943	0.86-0.988	0.920	0.746-0.993	0.930
NSC	Forward	Sex	A	0.988	0.988-0.988	0.987	0.987-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.988
			B	0.983	0.977-0.988	0.982	0.977-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.988
			C	0.471	0.419-0.547	0.454	0.398-0.531	0.725	0.488-0.988	0.701	0.46-0.987	0.619
		Cyto Risk	A	0.937	0.907-0.953	0.857	0.75-0.907	0.951	0.919-0.977	0.904	0.836-0.979	0.950
			B	0.892	0.837-0.93	0.860	0.707-0.933	0.891	0.872-0.907	0.873	0.801-0.919	0.891
			C	0.889	0.849-0.907	0.812	0.666-0.871	0.876	0.86-0.919	0.775	0.625-0.829	0.877
	Reverse	Sex	A	0.988	0.988-0.988	0.987	0.987-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.988
			B	0.526	0.453-0.547	0.487	0.439-0.52	0.504	0.436-0.547	0.481	0.426-0.52	0.515
			C	0.988	0.988-0.988	0.987	0.987-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.988
		Cyto Risk	A	0.926	0.872-0.977	0.828	0.656-0.986	0.929	0.907-0.977	0.913	0.846-0.986	0.927
			B	0.854	0.814-0.895	0.609	0.5-0.743	0.820	0.814-0.837	0.521	0.5-0.562	0.849
			C	0.907	0.884-0.925	0.907	0.863-0.954	0.886	0.872-0.942	0.852	0.794-0.964	0.905
SVM	Forward	Sex	A	0.980	0.965-0.988	0.980	0.966-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.988
			B	0.981	0.942-0.988	0.981	0.945-0.987	0.984	0.977-0.988	0.983	0.974-0.987	0.984
			C	0.516	0.43-0.593	0.515	0.429-0.599	0.753	0.43-0.988	0.752	0.426-0.987	0.671
		Cyto Risk	A	0.939	0.895-0.988	0.879	0.774-0.993	0.957	0.895-0.988	0.919	0.774-0.993	0.956
			B	0.963	0.919-0.988	0.941	0.829-0.993	0.961	0.93-0.988	0.947	0.868-0.993	0.961
			C	0.919	0.884-0.953	0.901	0.832-0.964	0.920	0.86-0.965	0.888	0.736-0.979	0.920
	Reverse	Sex	A	0.981	0.965-0.988	0.980	0.964-0.987	0.988	0.988-0.988	0.987	0.987-0.987	0.981
			B	0.493	0.407-0.593	0.489	0.404-0.592	0.497	0.407-0.593	0.497	0.409-0.588	0.495
			C	0.982	0.977-0.988	0.982	0.977-0.987	0.988	0.977-0.988	0.987	0.974-0.987	0.983
		Cyto Risk	A	0.932	0.872-0.977	0.871	0.739-0.962	0.965	0.919-0.988	0.955	0.812-0.993	0.932
			B	0.845	0.756-0.919	0.778	0.608-0.895	0.869	0.779-0.942	0.787	0.583-0.925	0.847
			C	0.944	0.872-0.977	0.901	0.729-962	0.964	0.907-0.988	0.944	0.846-0.993	0.944

AUC: Area under ROC curve; CI: Confidence interval; 95% CI was calculated from the bootstrap estimation.
A*: A_EFS_All; B*: B_OS_All; C*: C_SEX_All; D*: D_FAV_All; E*: E_EFS_HR; F*: F_OS_HR; AUC: Area under ROC curve; CI: Confidence interval; 95% CI was calculated from the bootstrap estimation. The upper-right and lower-left regions are for the cross-platform prediction of the models (the training and validation samples were profiled with different platforms), while the upper-left and lower-right regions are for the intra-platform prediction of the models (both training and validation samples were profiled with the same platform).

Back to article page

ISSN: 1474-760X

Contact us

Submission enquiries: editorial@genomebiology.com
General enquiries: info@biomedcentral.com

Genome Biology

Contact us