Shortest-Path Network Analysis of PTM Site Changes in Non-Small Cell Lung Cancer
Mentor: Dr. Karen Ross, Department of Biochemistry and Molecular & Cellular Biology, Georgetown University Medical Center.
Date/Time: August 25, 2020 at 3:20pm
Abstract: Tyrosine kinase inhibitors (TKI) can be used as drugs to target Non-Small Cell Lung Cancers (NSCLC). Due to pathway crosstalk, inhibitors may be effective in the short-term, but ineffective in the long-term due to development of resistance. Post-translational modifications (PTM) play an essential role in pathway regulation. Identification and analysis of PTM regulated pathways and PTM alteration profiles upon drug perturbation can provide insights into drug resistance mechanisms.
Previous work has established a method to identify PTM clusters given a quantitative measurement of PTMs (phosphorylation, acetylation, and ubiquitination) from several cell lines and drug treatment conditions. Based on the quantitative mass-spectrometer analysis of PTM, a co-cluster correlation network (CCCN) was created. To combine the correlated PTMs with pathways, a protein-protein interaction (PPI) network from curated sources is filtered using the clustering information in the CCCN to generate a PPI network called the Cluster Filtered Network (CFN). Shortest-path network analysis builds a composite shortest paths network (sub-network of CFN) that connects proteins with at least one significantly changed PTM site/sites (assigned threshold) to a specific molecule (drug target).
For the given cell lines, we identified the PTMs that changed significantly by drug perturbation and generated shortest-path networks connecting these sites to known drug targets. We also identified the cancer-related genes (oncogene and tumor suppressor genes) to examine their PTM changes in response to drug perturbation across non-small cell lung cancer cell lines. The results are presented mainly via heatmap and Cytoscape networks.
As a result, in terms of shortest-path network edges, the replicates experiments cluster, as well as experiments with the same drug treatment, suggesting a drug-dependent PTM change pattern, despite the divergences between two experimental batches (SEPTM and TenCell). Cancer-related gene neighbors in shortest-path network show a consistent relationship with experimental literature, providing clues for the cancerous signaling transduction. This approach provides a concise version of how a drug would impact a network. By integrating the information of shortest-path network of different tyrosine kinase inhibitors, the effect of combined drug use may be evaluated.