Predicting Personalized Responses to HDACi and PROTAC Therapies: A Data-Driven Approach

Abstract

Histone deacetylase inhibitors (HDACi) and PROTACs are promising cancer therapies, but predicting
their efficacy across different cell lines remains challenging. This study explores developing predictive
models for HDACi and PROTAC responses using multi-omics data and machine learning. For HDACi, it
is hypothesized that integrating genomic, transcriptomic, and proteomic data will identify biomarkers that
predict drug-induced gene expression changes and response across cancer cell lines. The complex
modulation of histone and non-histone proteins by HDACi necessitates multi-omics approaches to
understand their effects on cellular pathways. For PROTACs, it is hypothesized that machine learning
models incorporating protein expression and degradation data will predict drug responses by identifying
key proteins involved in the degradation process. PROTACs target previously "undruggable" proteins
through the ubiquitin-proteasome system. This model aims to enhance the precision of drug response
predictions. By leveraging these data-driven approaches, this research seeks to optimize personalized
cancer therapies and improve treatment outcomes for HDACi and PROTAC-based therapies.