PMC

A) List of 28 genes with patient and gene level summaries of the different variants. B) Toplogical analysis of network associations based on known biology among the set of positive and negative RAD genes. C) Association between genes of the negative and positive RAD sets. Each symbol is a gene with colors corresponding to the clustering analysis within that gene. Circle symbols indicate genes shared within our dataset while squares indicate associated genes not in our dataset. Solid lines indicate direct interactions while dotted lines indicate indirect interactions. Note the higher frequency of direct interactions among negative RAD genes.

A) Top biological pathways molecular functions for every patient’s set of significant preferentially expressed alleles. B) Druggable targets in every patient. The bar plots show the extent of preferential allele expression and the list is sorted by the average of allelic expression from highest to lowest. Drugs in bold type are used in current anticancer treatment or undergoing clinical anticancer trials. C) Network plots of a selected pathway in every patient showing a druggable target.

We observed in three different patients (P1, P2, P3) alleles in TP53 where one allele is significantly preferentially expressed. A) The allele positions are shown on a schematic view of the TP53 gene (red line) for all three patients. B) Alignment and coverage plots of the reads spanning the preferentially expressed alleles for both exome (blue) and RNA (brown) sequencing data. The patient and site is indicated on each plot. The histograms indicate coverage across the site with the coverage indicated by the y-axis number. The histogram central stacked bar plot shows the number of reference allele reads in blue and the alternate allele reads in red. The alignment plot shows the individual reads spanning the preferentially expressed alleles. For the RNA sequencing data, the dashed line indicate gaps corresponding to a spliced region.

A) Scatter plots show the exome and RNA allele fraction for every patient and site in our study and for cells lines (FB, GC, OE, SK). Grey circles are variants with non-significant allelic differences while significant differences are indicated in colored circles as stated in the Legend. Note we imposed a cutoff of allelic difference of 0.2. B) The number of significant non-synonymous coding variants relative to all variants for patients and cell lines. C) The shared variants in patients and cell lines are shown as bar plots. D) Analysis of the significance of shared variants. The approximate p-value for the pair is shown in the heatmap with p-values of 0 being black and p-values of 1 being white. We analyzed the positive RAD variants and the negative RAD variants and found large difference between the cell line data and the patient site.

Clustering of the allele fraction differences for all somatic genes shows somatic mutations with preferential allele expression for the mutant allele (orange), no preferential expression (white), reference allele preferential expression (blue) and no expression of any allele (gray). Subplots for each patient show the exome (blue circles) and RNA (red circles) allele fraction for every site (NO is normal, OV is ovary, PE is peritoneum and LN is lymph node). Filled blue circles indicate significant difference from 1 and filled red circles indicate significant difference from the exome allele fraction for the given sites. Open circles indicate no significant difference. An absence of a circle indicates no data for that site. Plots are sorted from the highest allele fraction difference to the lowest. Genes with no RNA expression were arbitrarily assigned a highly negative allele fraction difference. Shaded plots indicate if there’s significant preferential allele expression and the direction of the expression as stated in the legend. Note that shading was done as an aggregate of all sites within a patient. For example, in P2, one site in PIK3CA has a significant allelic difference (the LN) while there is no difference at other sites.

A) Overlap of preferentially expressed variant alleles for the different sampled sites (OV, PE, LN). Numbers in intersecting or non-intersecting regions indicate the number of shared genes. B) Ingenuity pathway analysis of the different genes in each site showing the top 3 predicted diseases and the biological functions for each site. Note that gastrointestinal disease indicates cancer of the gastrointestinal system. Threshold (orange line) is set at a p-value of 0.05. C) Hierarchical clustering analysis of RNA-exome allele fraction differences expression differences show that different sites have largely different allelic expression patterns. As in previous figures, a positive allelic difference indicates a reference allele is preferentially expressed while a negative allelic difference indicates an alternate allele is preferentially expressed. Patients are also different with few shared alleles. D) Parallel plots of allele fraction differences for different patients and different site expression patterns. The site or combination of sites where there is a negative allele fraction difference is indicated by the site names (OV, PE, LN) for ovary, peritoneum and lymph node. Gene names that appear in at least two patients for a given pattern are indicated to the right of the parallel plot.