Additional file 1: of Imprinted and ancient gene: a potential mediator of cancer cell survival during tryptophan deprivation

Performance of 4 normalisation methods and 2 transcript abundance estimation procedures on the TOIL TCGA/GTEX dataset used in this study. A) Pairwise correlations of 9 normalisation/quantitation methods. Each circle represents an arithmetic mean of 30 replicated Spearman’s correlation values each calculated between 1 × 107 randomly sampled gene expression pairs from each dataset (each dataset contains 3 × 107 expression values; 19,446 genes × 15,741 samples). B) The effect of 5 normalisation techniques on differential expression (t-statistic) of 12 genes (used in Figs. 2 and 3 of the main manuscript) between each of the 28 cancer types and their respective normal tissues. t-statistic values were calculated using Welch’s unequal variance t-test. C) Distribution of IMPACT, IDO1 and TDO2 expression values across all 15,741 non-cancer (GTEX) and cancer (TCGA) samples examined in this study. D) Performance of 5 normalisation methods in assessing differential expression (t-statistic) of TDO2, IDO1 and IMPACT in 28 cancer types relative to their corresponding normal tissues. (PDF 576 kb