Explainable Machine Learning and Deep Learning Models for Predicting TAS2R-Bitter Molecule Interactions
Authors:
Francesco Ferri,
Marco Cannariato,
Lorenzo Pallante,
Eric A. Zizzi,
Marco A. Deriu
Abstract:
This work aims to develop explainable models to predict the interactions between bitter molecules and TAS2Rs via traditional machine-learning and deep-learning methods starting from experimentally validated data. Bitterness is one of the five basic taste modalities that can be perceived by humans and other mammals. It is mediated by a family of G protein-coupled receptors (GPCRs), namely taste rec…
▽ More
This work aims to develop explainable models to predict the interactions between bitter molecules and TAS2Rs via traditional machine-learning and deep-learning methods starting from experimentally validated data. Bitterness is one of the five basic taste modalities that can be perceived by humans and other mammals. It is mediated by a family of G protein-coupled receptors (GPCRs), namely taste receptor type 2 (TAS2R) or bitter taste receptors. Furthermore, TAS2Rs participate in numerous functions beyond the gustatory system and have implications for various diseases due to their expression in various extra-oral tissues. For this reason, predicting the specific ligand-TAS2Rs interactions can be useful not only in the field of taste perception but also in the broader context of drug design. Considering that in-vitro screening of potential TAS2R ligands is expensive and time-consuming, machine learning (ML) and deep learning (DL) emerged as powerful tools to assist in the selection of ligands and targets for experimental studies and enhance our understanding of bitter receptor roles. In this context, ML and DL models developed in this work are both characterized by high performance and easy applicability. Furthermore, they can be synergistically integrated to enhance model explainability and facilitate the interpretation of results. Hence, the presented models promote a comprehensive understanding of the molecular characteristics of bitter compounds and the design of novel bitterants tailored to target specific TAS2Rs of interest.
△ Less
Submitted 21 June, 2024;
originally announced June 2024.
LightCPPgen: An Explainable Machine Learning Pipeline for Rational Design of Cell Penetrating Peptides
Authors:
Gabriele Maroni,
Filip Stojceski,
Lorenzo Pallante,
Marco A. Deriu,
Dario Piga,
Gianvito Grasso
Abstract:
Cell-penetrating peptides (CPPs) are powerful vectors for the intracellular delivery of a diverse array of therapeutic molecules. Despite their potential, the rational design of CPPs remains a challenging task that often requires extensive experimental efforts and iterations. In this study, we introduce an innovative approach for the de novo design of CPPs, leveraging the strengths of machine lear…
▽ More
Cell-penetrating peptides (CPPs) are powerful vectors for the intracellular delivery of a diverse array of therapeutic molecules. Despite their potential, the rational design of CPPs remains a challenging task that often requires extensive experimental efforts and iterations. In this study, we introduce an innovative approach for the de novo design of CPPs, leveraging the strengths of machine learning (ML) and optimization algorithms. Our strategy, named LightCPPgen, integrates a LightGBM-based predictive model with a genetic algorithm (GA), enabling the systematic generation and optimization of CPP sequences. At the core of our methodology is the development of an accurate, efficient, and interpretable predictive model, which utilizes 20 explainable features to shed light on the critical factors influencing CPP translocation capacity. The CPP predictive model works synergistically with an optimization algorithm, which is tuned to enhance computational efficiency while maintaining optimization performance. The GA solutions specifically target the candidate sequences' penetrability score, while trying to maximize similarity with the original non-penetrating peptide in order to retain its original biological and physicochemical properties. By prioritizing the synthesis of only the most promising CPP candidates, LightCPPgen can drastically reduce the time and cost associated with wet lab experiments. In summary, our research makes a substantial contribution to the field of CPP design, offering a robust framework that combines ML and optimization techniques to facilitate the rational design of penetrating peptides, by enhancing the explainability and interpretability of the design process.
△ Less
Submitted 31 May, 2024;
originally announced June 2024.
