Fig. 4

Methodology of model annotation with patient-specific omics data followed by dynamical analysis (Step 3 of CanSeer). (A) The rules-based model is customized with patient-specific omics data including gene expressions, copy number variations (CNVs), somatic mutations (SMs) and genomic structural variants (SVs). The model incorporating patient-specific data is analyzed through dynamical analysis, and the resulting cell fate outcomes are compared between normal and cancer conditions. (B) In normal tissue samples categorized as paired case, normalized gene expression (NGE) values are assigned to the input nodes, and output node propensities are compared to corresponding gene expressions, with network tuning if needed. For unpaired cases, the median NGE values are used for comparison and tuning. (C) In the network model, cancer samples, whether they belong to paired, unpaired, or cancer-only cases, are annotated in the same way. First, patient-specific altered genes (CNVs, SMs and SVs) are searched within the network. Then, the types of SMs are considered for further inclusion and exclusion. SMs such as missense mutation, splice donor variant, splice acceptor variant, splice site variant, frameshift deletion, and non-coding transcript exon variant are included while silent mutation, intron variant, splice region variant, and 5ʹ UTR variant are excluded. Next, the patient-specific altered genes are mapped with the driver genes of same TCGA project and cancer type. The identified patient-specific CNVs, SMs, and SVs are fixed into the model along with the input nodes. These nodes are fixed with the corresponding NGEs of the patient’s cancer sample, to maintain their activity constant throughout the analysis, while other network nodes remain dynamic. The output node propensities from dynamical analysis are compared with the corresponding patient NGEs of cancer sample. If the propensities misalign with the respective expression values, the network is tuned. (D) The cell fates (apoptosis, cell cycle arrest, proliferation, and metastasis) are compared between patient-specific normal and cancer conditions in paired case, and median assigned normal and patient-specific cancer conditions in unpaired case. In cancer only case, only patient-specific cancer fate propensities are plotted. In the bar chart, the x-axis represents cell fates, and the y-axis represents propensities.