Skip to main content
Fig. 3 | BMC Cancer

Fig. 3

From: Contribution and clinical relevance of germline variation to the cancer transcriptome

Fig. 3

Detection of eQTLs in cancer tissue. A Correlation between somatic alteration rates and eGene fractions in the downsampled dataset. (Left) Correlation between tumor mutation burden and eGene fraction. (Right) Copy number alteration rate was measured by the fraction of genome altered by any type of copy number alteration. Spearman’s correlation statistics are shown. B Correlation between eGene fraction when considering all genes (x-axis) and when considering quiet genes only (those altered in subsets of samples) in the downsampled dataset. (Left) Quiet genes defined as those with no somatic alteration in any sample. (Middle) Quiet genes defined as those with somatic alterations in at most 5% of samples. (Right) Quiet genes defined as those with somatic alterations in at most 20% of samples. Only genes expressed in all samples were used in this analysis. Spearman’s correlation statistics are shown. C Correlations between tumor microenvironment scores as defined by ESTIMATE and eGene fraction. Standardized immune (left) and stromal (right) scores defined by ESTIMATE were obtained for all cancer types using the downsampled data, and the median score is shown for each cancer type. Spearman’s correlation statistics are shown. D Correlation between levels of intratumor heterogeneity and eGene fraction in downsampled data. The mutant allele tumor heterogeneity (MATH) score was computed for every sample (Methods) and the median MATH score was used to summarize intratumor heterogeneity for each cancer type. Higher MATH scores indicate higher levels of intratumor heterogeneity. Spearman’s correlation statistics are shown

Back to article page