GuacaMol: Benchmarking Models for De Novo Molecular Design

By Nathan Brown, Marco Fiscato, Marwin H. S. Segler, Alain C. Vaucher. 2018

942 citations recorded by the source corpus.

De novo design seeks to generate molecules with required property profiles by virtual design-make-test cycles. With the emergence of deep learning and neural generative models in many application areas, models for molecular design based on neural networks appeared recently and show promising results. However, the new models have not been profiled on consistent tasks, and comparative studies to well-established algorithms have only seldom been performed. To standardize the assessment of both classical and neural models for de novo molecular design, we propose an evaluation framework, GuacaMol, based on a suite of standardized benchmarks. The benchmark tasks encompass measuring the fidelity of the models to reproduce the property distribution of the training sets, the ability to generate novel molecules, the exploration and exploitation of chemical space, and a variety of single and multiobjective optimization tasks. The benchmarking open-source Python code and a leaderboard can be found on https://benevolent.ai/guacamol .

Subfields: Computational Theory and Mathematics, Environmental Chemistry, Materials Chemistry

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Environmental Chemistry

In Materials Chemistry

Citations in the corpus

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns (2019, background): “…frameworks based on a variety of generative models for modeling molecules, which diﬀer in the respective model structure and in the selected input features (Kadurin et al., 2017; Gupta et al., 2018; Jørgensen et al., 2018b; Arús-Pous et al., 2019; Brown et al., 2019; Jensen, 2019; Xu et al., 2019).”
Artificial intelligence in drug discovery: recent advances and future perspectives (2021, methodology): “Recent efforts include the MOSES [154] and GuacaMol platforms [155], which implement several popular neural generative models, as well as more classical models ( e.g . genetic algorithms [156]), and provide several metrics for their comparison.”
Differentiable Scaffolding Tree for Molecular Optimization (2021, positive): “Also, as they learn the reference data distribution, their ability to explore diverse chemical space is relatively limited, evidenced by the recent molecular optimization benchmarks [4, 18].”
Graph convolutional networks for computational drug development and discovery (2019, methodology): “[110] introduced an evaluation framework, GuacaMol, based on a suite of standardized benchmarks.”
Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development (2021, methodology): “– FCD Score (Fréchet ChemNet Distance) first takes the means and covariances of activations of the penultimate layer of ChemNet as calculated for the reference set and for the set of generated molecules (Brown et al. 2019, Preuer, Renz, Unterthiner, Hochreiter & Klambauer 2018).”
Randomized SMILES strings improve the quality of molecular generative models (2019, methodology): “For this reason, some methods were published [36, 40] that aggregate a set of metrics to obtain a better overview of the output domain of the model.”