Identification of protein properties associated with prostate cancer EVs using machine learning

Abstract

In this study, we focus on proteins released by prostate cancer cells that form small particles called extracellular vesicles (EVs). EVs play a crucial role in how cancer spreads in the body by carrying a variety of biomolecules. Proteins are one of these molecules that hold promise as potential biomarkers to help us understand and utilize EVs. We used a computational technique called machine learning to predict the types of proteins come from prostate cancer cells by incorporating 95 features of the proteins. The results showed that the machine learning program could predict the proteins in these EVs with moderate accuracy. When we paid special attention to the most common proteins found in these EVs, the predictions became even better. This study also revealed that certain characteristics of these proteins, such as palmitoylation, played an important role in the predictions. Overall, this study demonstrates that using machine learning to understand the proteins in these cancer-related EVs can provide valuable insights for finding new ways to diagnose and treat cancer.