MolFinder: an evolutionary algorithm for the global optimization of molecular properties and the extensive exploration of chemical space using SMILES

Abstract

Here, we introduce a new molecule optimization method, MolFinder, based on an efficient global optimization algorithm, the conformational space annealing algorithm, and the SMILES representation. MolFinder finds diverse molecules with desired properties efficiently without any training and a large molecular database. Compared with recently proposed reinforcement-learning-based molecule optimization algorithms, MolFinder consistently outperforms in terms of both the optimization of a given target property and the generation of a set of diverse and novel molecules. The efficiency of MolFinder demonstrates that combinatorial optimization using the SMILES representation is a promising approach for molecule optimization, which has not been well investigated despite its simplicity. We believe that our results shed light on new possibilities for advances in molecule optimization methods.

Scalable estimator of the diversity for de novo molecular generation resulting in a more robust QM dataset (OD9) and a more efficient molecular optimization

“…We believe that the inclusion of a diversity criterion can improve the interest of the solutions proposed by any generator. During the course of this study, Kwon et al published an article where they add a criterion based on the Tanimoto similarity on fingerprints into their evolutionary algorithm [ 24 ]. This article confirms our opinion because MolFinder maintains great optimization performances in a reference benchmark despite this additional diversity criterion.…”

V-Dock: Fast Generation of Novel Drug-like Molecules Using Machine-Learning-Based Docking Score and Molecular Optimization

“…This work aims to design drug-like molecules with desirable properties by combining docking score prediction models and the global molecular property optimization approach, MolFinder [ 2 ]. MolFinder performs global optimization of a given objective function through combinatorial optimization of SMILES based on the conformational space annealing (CSA) algorithm.…”

“…In this work, to overcome this limitation and facilitate the computational drug design process, we propose a new computational workflow. V-dock uses machine-learning algorithms to predict the docking score, followed by the application of the global molecular property optimization algorithm, MolFinder, to optimize the score [ 2 ]. We show that this new workflow significantly reduces the required computational resources compared to conventional approaches based on ligand docking and facilitates the design of novel drug-like molecules with the desired properties.…”

“…The function consists of three terms: predicted docking score, drug-likeness score (QED), and similarity to a reference molecule. The MolFinder [ 2 ] algorithm, the global molecular property optimization algorithm using SMILES, was used to optimize the given objective function. Previously, we demonstrated that MolFinder generates more optimized molecules than known RL-based methods.…”

“…The number of known inhibitors of WRN is limited, making it a challenging target for existing reinforcement-based models, which heavily depend on large amounts of biological activity data. Compared to existing generative models, MolFinder has the advantage of generating diverse molecules with desirable properties without many experimental data because it samples novel molecules in a combinatorial manner in the SMILES space [ 2 ]. In addition, we identified that our approach, which uses a docking score predictor, accelerates the molecular optimization process 255.5 times compared to actual docking calculations.…”