Abstract LB060: Integrated analysis of transcriptomics and kinase activity data for better characterization of cancer models

Abstract Introduction: Quantitative measurements of transcripts and proteins are key to investigate the basal state of a biological system, while functional proteomics inform about the active state of regulatory networks. Here we describe how the integration of transcriptomics and kinase activity data lead to a better characterization of various cancer models. Methods: We performed RNA sequencing (RNAseq) and kinase activity profiling of 63 Patient Derived Xenograft (PDX) models from six tumor types (Breast, Ovarian, Colon, Melanoma, Lung and Acute Myeloid Leukemia, AML). RNAseq was performed on an Illumina NovaSeq platform. The data was DESeq2-normalized and log2-transformed. Protein Tyrosine Kinase and Serine-Threonine Kinase activities were profiled on PamChip® peptide microarray. To identify the role of kinase signaling related genes we defined a set of signaling-specific genes (n=2932), based on the elements from the reactome signal transduction pathway database (n=2560) and additional kinases (n=372) represented on Pamchip, that was used for further analysis. Integrated analysis of transcriptomics and kinase activity data was performed using Multi Omics Factor Analysis (MOFA). Results: Principal Component Analysis (PCA) of RNAseq data using all included genes or 2932 kinase signaling-specific genes showed clustering of the data according to cancer type, with ovarian cancer showing most heterogeneity, which indicates the importance of kinase signaling in these malignancies. Interestingly, with integrated RNAseq-Kinase activity data all except ovarian cancer show clustering of cancer types on the MOFA Factor 1 - Factor 3. Pathway analysis on the highest ranking 100 genes from principal component 1 of RNAseq data (capturing variation between AML and the other tumor types) resulted in 60 KEGG pathways. Importantly, highest ranking 50 genes and 47 peptides comprising MOFA Factor1 identified 115 significant KEGG pathways, and the statistical score of pathways identified by RNAseq alone was further improved. Finally, significant correlation between gene expression and kinase activity was found for selected PDX model per malignancy. Furthermore, ranking PDX models based on correlation score provided suitable tool to select PDX models for disease or pathway specific research question. Conclusion: Integrating transcriptomics with kinase activity data can be used to confirm transcriptomics findings on a functional level and provides deeper biological insights than transcriptomics alone. We show that integrative analysis leads to more significant and a higher number of enriched pathways. High correlation between two datasets allows for selecting animal models addressing specific research questions. Integrated analysis of transcriptomics and kinase activity data has great potential in improving diagnosis, prognosis and prediction of response to treatment. Citation Format: Dóra Schuller, Rik de Wijn, Dirk Pijnenburg, Tobias Deigner, Julia Schueler, Simar Pal Singh. Integrated analysis of transcriptomics and kinase activity data for better characterization of cancer models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB060.