Fig. 9
From: DegNorm: normalization of generalized transcript degradation improves accuracy in RNA-seq analysis
Comparison of different normalization methods in simulations II, III, and IV. a–d Scatter plot of normalized read count vs. latent read count in log2 scale with diagonal line imposed from DegNorm, UQ, RUVg, RUVr, and TIN methods for simulation II. e–g Coefficient of variation (CV) vs. normalized mean read count (in log scale) for 18,000 true-negative genes (out of a total of 20,000 genes). The bean pot under the CV plot shows the density of log of mean read counts from DegNorm. h–j Empirical CDF of p value from edgeR DE analysis. The latent curve corresponds to the results using the true read counts before the degradation was imposed. k–m Receiver operating characteristic curves of DE analysis