A normalization strategy for comparing tag count data

Table 1 Average AUC values for two edgeR-related combinations.

	P_A = 50%	60%	70%	80%	90%	100%
	(a) edgeR/TbT
P_DEG = 5%	90.52	89.92	90.58	90.67	90.59	90.10
10%	90.33	90.23	90.80	90.14*	91.02*	90.39*
20%	90.43	90.53	90.52*	90.60*	90.41*	90.41*
30%	90.71	90.66*	90.23*	90.67*	90.00*	89.46*
	(b) edgeR/default
5%	90.52	89.92	90.56	90.62	90.50	89.95
10%	90.33	90.21	90.73	89.99	90.74	89.89
20%	90.43	90.49	90.26	90.00	89.24	88.40
30%	90.71	90.54	89.58	89.35	87.20	84.55

Average AUC values of total of 100 trials for each simulation condition: (a) edgeR/TbT and (b) edgeR/default. Simulation data contain a total of 20,000 genes: P_DEG % of genes is DEGs, and P_A % of P_DEG is higher in Sample A. A total of 24 conditions (four P_DEG values × six P_A values) are shown. Highest AUC value for each condition is in bold. AUC values with asterisks indicate significant improvements (p-value < 0.01, Wilcoxon rank sum test).

ISSN: 1748-7188