Fig. 2

Expression outliers are enriched for hematologic malignancy driver genes. A Enrichment for CGC hematologic tumor suppressor genes among all genes called by OUTRIDER as well as at most three significant genes per sample that were called to be underexpression outliers. The genes are stratified by the number of samples in which the gene is called as an outlier. Numbers of the genes and nominal significances from the Fisher test are labeled at the top of the bars (ns: not significant; *: P ≤ 0.05; **: P ≤ 0.01; ***: P ≤ 0.001; ****: P ≤ 0.0001). Error bars of the odds ratio (95% confidence intervals) are shown where the Fisher test is significant. B, C As in A for CGC hematologic oncogenes among overexpression and activation outliers, respectively. D Raw RNA-seq read counts per sample against expected counts for TET2. For fifteen samples (red), TET2 was called as an underexpression outlier, including samples AML_0405 and T-ALL_0269, showcased in the panels (E–H). E, F Gene expression z-score among all samples against the genomic position of the genes on chromosome 4. G, H Copy ratio within chromosome 4 binned into 1-kb region. Red vertical lines mark the genomic position of the expression outliers. Black dashed lines mark the estimated region of no copy number variation. The underexpression outliers in TET2 and UBE2D3 in sample AML_0405 reflect the consequence of copy number loss that is very specific to the TET2 locus. In contrast, an extra copy of whole chromosome 4 (karyotype: 47,XY, + 4) is found in sample T-ALL_0269. TET2 reduced expression must have a different cause in this sample