Leveraging Explainable AI for Improved Understanding and Prediction of Drug Responses in IGF1R Signaling Pathways

Pharmacogenomics showcases the aim of precision medicine, which strives to customize treatments for individuals and specific populations. This field delves into exploring how an individuals DNA influences their response to medications. A persons genetic composition can impact the likelihood of experiencing reactions or determining the effectiveness of a medication. By providing insights into the safety and effectiveness of drug therapies pharmacogenomics holds potential for significantly enhancing health outcomes. Through advancements in targeted therapies we can precisely target abnormalities that trigger tumor growth in patients. For instance IGF1R (Insulin like Growth Factor 1 Receptor) which belongs to the tyrosine kinase receptor family plays a crucial role in promoting cell growth, survival and proliferation across different types of cancers. The overexpression of IGF1R has been observed in cancer types indicating its involvement in fueling continuous growth and survival of cancer cells. Targeting IGF1R helps address the dysregulation of this receptor within cancer cells. Artificial Intelligence (AI) comes into play by enabling prediction of suitable drugs based on a patients genomic profile thereby reducing adverse effects and improving treatment effectiveness. Parallel, here has been growing concern regarding model explanation due, to the opaque nature of model predictions. This is particularly important when it comes to modeling drug responses. In our research paper we have employed AI to gain a clear understanding of the prediction model and the factors that affect its results. The findings show that lower valued counts of YAPₚS127Caution protein tend to negatively impact the output. Similarly lower values of YAPₚS127Caution protein and higher valued counts of YAPₚS127 Caution protein, Xanthine, Tyrosine tends to positively impact the output. This helps as an aiding reference in knowing which feature of an unknown cell line should be focused to know drug response prior medication of BMS-536924, BMS-754807 drug.