MARS: Markov Molecular Sampling for Multi-objective Drug Discovery

Xie, Yutong; Shi, Chence; Zhou, Hao; Yang, Yuwei; Zhang, Weinan; Yu, Yong; Li, Lei

doi:10.48550/arxiv.2103.10432

Cited by 23 publications

(49 citation statements)

References 28 publications

(30 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Note that the baseline methods we compare to do not constrain synthesizability, which means they explore a larger chemical space and are able to obtain molecules that score higher but are not synthesizable. The results show that our model consistently outperforms GCPN [12] and MolDQN [48], and is comparable to GA+D [49] and MARS [50] across different tasks. We highlight the case of GSK3β inhibitor optimization in Figure 6.…”

Section: Synthesizable Molecular Optimization Resultsmentioning

confidence: 69%

Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design

Gao¹,

Mercado²,

Coley³

2021

Preprint

View full text Add to dashboard Cite

Molecular design and synthesis planning are two critical steps in the process of molecular discovery that we propose to formulate as a single shared task of conditional synthetic pathway generation. We report an amortized approach to generate synthetic pathways as a Markov decision process conditioned on a target molecular embedding. This approach allows us to conduct synthesis planning in a bottom-up manner and design synthesizable molecules by decoding from optimized conditional codes, demonstrating the potential to solve both problems of design and synthesis simultaneously. The approach leverages neural networks to probabilistically model the synthetic trees, one reaction step at a time, according to reactivity rules encoded in a discrete action space of reaction templates. We train these networks on hundreds of thousands of artificial pathways generated from a pool of purchasable compounds and a list of expert-curated templates. We validate our method with (a) the recovery of molecules using conditional generation, (b) the identification of synthesizable structural analogs, and (c) the optimization of molecular structures given oracle functions relevant to drug discovery.

show abstract

Section: Synthesizable Molecular Optimization Resultsmentioning

confidence: 69%

Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design

Gao¹,

Mercado²,

Coley³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Particular implementations include ChemTS [43] that relies on string-based representations of molecular graphs and an RNN for modeling the MCTS. Alternative approaches are unitMCTS [44] and MARS [45], which both rely on matrix representations of molecular graphs. While unitMCTS implements MCTS for modeling the molecular generation process, MARS uses annealed Markov chain Monte Carlo sampling.…”

Section: Deep Generative Modelsmentioning

confidence: 99%

“…The corresponding structures are depicted in Figure 3. In recent years, it has been claimed that the unconstrained penalized log P maximization task is not useful for benchmarking inverse molecular design algorithms as it has trivial solutions simply producing molecules with ever longer chains [45]. However, that is only true if no SMILES character limit is enforced.…”

Section: Unconstrained Penalized Log P Optimizationmentioning

confidence: 99%

JANUS: Parallel Tempered Genetic Algorithm Guided by Deep Neural Networks for Inverse Molecular Design

Nigam,

Pollice,

Aspuru-Guzik

2021

Preprint

View full text Add to dashboard Cite

Inverse molecular design, i.e., designing molecules with specific target properties, can be posed as an optimization problem. High-dimensional optimization tasks in the natural sciences are commonly tackled via population-based metaheuristic optimization algorithms such as evolutionary algorithms. However, expensive property evaluation, which is often required, can limit the widespread use of such approaches as the associated cost can become prohibitive. Herein, we present JANUS, a genetic algorithm that is inspired by parallel tempering. It propagates two populations, one for exploration and another for exploitation, improving optimization by reducing expensive property evaluations. Additionally, JANUS is augmented by a deep neural network that approximates molecular properties via active learning for enhanced sampling of the chemical space. Our method uses the SELFIES molecular representation and the STONED algorithm for the efficient generation of structures, and outperforms other generative models in common inverse molecular design tasks achieving state-of-the-art performance.Preprint. Under review.

show abstract

“…Most similar to our work, we highlight two recent models based on Markov Chain Monte Carlo (MCMC). The MARS model [Xie et al, 2021] trains and generates molecules simultaneously, adding better candidates to its training set. The generation procedure shares a lot of similarity with our work.…”

Section: Optimizationmentioning

confidence: 99%

Realistic molecule optimization on a learned graph manifold

Brossard,

Frigo,

Dehaene

2021

Preprint

View full text Add to dashboard Cite

Deep learning based molecular graph generation and optimization has recently been attracting attention due to its great potential for de novo drug design. On the one hand, recent models are able to efficiently learn a given graph distribution, and many approaches have proven very effective to produce a molecule that maximizes a given score. On the other hand, it was shown by previous studies that generated optimized molecules are often unrealistic, even with the inclusion of mechanics to enforce similarity to a dataset of real drug molecules. In this work we use a hybrid approach, where the dataset distribution is learned using an autoregressive model while the score optimization is done using the Metropolis algorithm, biased toward the learned distribution. We show that the resulting method, that we call learned realism sampling (LRS), produces empirically more realistic molecules and outperforms all recent baselines in the task of molecule optimization with similarity constraints.Preprint. Under review.

show abstract

MARS: Markov Molecular Sampling for Multi-objective Drug Discovery

Cited by 23 publications

References 28 publications

Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design

Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design

JANUS: Parallel Tempered Genetic Algorithm Guided by Deep Neural Networks for Inverse Molecular Design

Realistic molecule optimization on a learned graph manifold

Contact Info

Product

Resources

About