(M1430-02-12) Integrating Omics Data Through AI: Predicting Novel Therapeutic Targets for Therapy-Resistant Cancer Patients

Purpose: Artificial intelligence (AI) is transforming biomedicine by facilitating the thorough analysis of multi-omics data, thereby deepening our grasp of intricate biological systems and the underlying mechanisms of diseases. Our study leverages AI to synthesize diverse omics datasets—including genomics, proteomics, and metabolomics—with the goal of identifying novel therapeutic targets for cancer patients resistant to current treatments. This strategy is designed to enhance the development of personalized medicine and refine treatment approaches, offering new avenues for addressing complex medical challenges.

Methods: Biopsy samples from cancer patients who are resistant to current therapies were collected to produce extensive metabolomics, proteomics, and transcriptomics datasets. These datasets shed light on the metabolic pathways, protein functions, and gene expression patterns linked to therapy resistance. We integrated this multi-omics data using sophisticated computational techniques to construct detailed molecular profiles for each sample. Subsequently, an unsupervised AI model was developed to scrutinize these profiles, identifying patterns and correlations that traditional methods might miss. This AI model enabled us to formulate hypotheses about new potential drug targets, particularly targeting the mechanisms and pathways most closely associated with resistance.

Results: The in-depth analysis of biopsy samples from cancer patients resistant to immunotherapies provided deeper insights into the resistance mechanisms, uncovering intricate details that could guide the development of new therapeutic interventions. Our metabolomics investigations pinpointed specific disruptions in metabolic pathways that are crucial for the energy supply and survival of resistant cancer cells, highlighting potential targets for metabolic inhibitors. Proteomics analysis revealed that protein alterations involved not only basic regulatory functions but also complex mechanisms of immune evasion, such as the suppression of antigen presentation and the modulation of checkpoint proteins. Transcriptomics data illuminated extensive gene expression changes that influence both local immune responses and broader interactions within the tumor microenvironment, such as cytokine signaling and cell adhesion processes. Through AI modeling, we synthesized these findings to identify novel, interconnected patterns across the datasets, leading to the identification of promising new drug targets. These targets are involved in key processes and pathways that sustain the resistance phenotype, offering fresh avenues for therapeutic exploration to overcome resistance in immunotherapy.

Conclusion: Our study underscores the substantial impact of integrating multi-omics data with AI analysis in uncovering the complex resistance mechanisms of cancer to immunotherapies, identifying promising new drug targets. By systematically examining metabolomics, proteomics, and transcriptomics datasets from biopsy samples of therapy-resistant cancer patients, we identified significant metabolic disruptions, alterations in protein regulation, and gene expression changes that are crucial for the persistence and survival of these resistant cells. These insights have enabled us to identify novel drug targets within these resistance pathways, focusing on critical and previously uncharted mechanisms. Integrating Omics Data through AI: Predicting Novel Therapeutic Targets for Therapy-Resistant Cancer Patient; Integrating omics data though AI; Predicting novel therapeutic targets for therapy-resistant cancer patients.