Multivariate models from RNA-Seq SNVs yield candidate molecular targets for biomarker discovery: SNV-DA

BMC Genomics

Table 4 Breast model performances

Model	Tested Range of K, Every Nth	Opt. K	AUC [95 % CI]	P-value	Pred. Accuracy	ER+ Sens.	TR- Sens.
intergenic	10–1000, 40	771	0.975 [0.972–0.977]	<0.001	0.939	0.922	0.956
all-SNVs	10–1000, 40	163	0.972 [0.969–0.975]	<0.001	0.941	0.936	0.968
up/downstream	10–400, 15	386	0.960 [0.959–0.962]	<0.001	0.915	0.910	0.920
exonic	10–1000, 40	50	0.958 [0.952–0.964]	<0.001	0.912	0.906	0.917
3’ UTR	10–1000, 40	129	0.939 [0.936–0.942]	<0.001	0.884	0.914	0.854
ncRNA	10–1000, 40	911	0.939 [0.936–0.942]	<0.001	0.843	0.917	0.768
5’ UTR	10–400, 15	370	0.939 [0.931–0.946]	<0.001	0.837	0.873	0.801
intronic	10–1000, 40	315	0.935 [0.933–0.937]	<0.001	0.879	0.829	0.930
nonsynonymous exonic	10–1000, 40	92	0.920 [0.915–0.926]	<0.001	0.869	0.857	0.881
RNA-editing	10–200, 5	12	0.878 [0.873–0.883]	<0.001	0.820	0.747	0.894
gene expression	10–1000, 40	472	0.985 [0.983–0.987]	<0.001	0.963	0.976	0.951

The range of K tested, the optimal value of K, AUC and 95 % confidence interval, p value from 1000 iteration permutation tests, predictive accuracy, and classification sensitivities of the top-performing models by genomic region

ISSN: 1471-2164