Deep Modeling of Regulating Effects of Small Molecules on Longevity-Associated Genes

Area of Science:

• Computational biology and bioinformatics
• Genomics and computational drug discovery
• Aging research and longevity science

Background:

• Aging is a natural process causing cellular and organ dysfunction.
• Large-scale gene expression data and machine learning advance drug repurposing for longevity.
• Previous work introduced DeepCOP, a quantitative chemogenomic model for predicting gene-regulating effects.

Purpose of the Study:

• To extend the DeepCOP model for predicting aging gene-regulating effects of small molecules across multiple cell lines.
• To develop a quantitative chemogenomic deep model for predicting gene expression responses to chemical perturbations.
• To identify drugs that positively influence aging-related genes and promote longevity.

Main Methods:

• Developed and applied an extended quantitative chemogenomic model (Deep Model) using gene ontology labels, molecular fingerprints, and cell line descriptors.
• Trained the deep model to predict gene expression responses to chemical perturbations across various cell lines (LINCS dataset).
• Evaluated state-of-the-art machine learning approaches, including deep neural network (DNN) classifiers, as benchmarks.

Main Results:

• The deep neural network (DNN) classifier successfully ranked known drugs with beneficial effects on aging genes.
• Several identified drugs were previously demonstrated to promote longevity, validating the model's predictive capability.
• Demonstrated the effectiveness of 'hybrid' chemogenomic models integrating quantitative biomarker descriptors for cell-specific drug-gene interactions.

Conclusions:

• The developed quantitative chemogenomic deep model accurately predicts gene expression responses to chemical perturbations.
• This methodology holds significant potential for discovering novel drugs targeting aging and promoting longevity.
• Hybrid chemogenomic models offer a powerful approach for understanding and manipulating cell-specific drug-gene interactions in aging research.