Generating Focused Molecule Libraries for Drug Discovery with Recurrent Neural Networks

@article{Segler2017GeneratingFM,
  title={Generating Focused Molecule Libraries for Drug Discovery with Recurrent Neural Networks},
  author={Marwin H. S. Segler and Thierry Kogej and Christian Tyrchan and Mark P. Waller},
  journal={ACS Central Science},
  year={2017},
  volume={4},
  pages={120 - 131}
}
In de novo drug design, computational strategies are used to generate novel molecules with good affinity to the desired biological target. [] Key Result When coupled with a scoring function, our model can perform the complete de novo drug design cycle to generate large sets of novel molecules for drug discovery.
View on PubMed
pubs.acs.org
839 Citations
Highly Influential Citations
35
Background Citations
166
Methods Citations
150
Results Citations
5

839 Citations

In silico generation of novel, drug-like chemical matter using the LSTM neural network

A method to generate molecules using a long short-term memory (LSTM) neural network and an analysis of the results, including a virtual screening test, confirms that the potential of these novel molecules to show bioactivity is comparable to the ChEMBL set from which they were derived.

Training recurrent neural networks as generative neural networks for molecular structures: how does it impact drug discovery?

An overview of recurrent neural network-based generative models and their variants for molecule generation with recent examples and models that could be trained on molecular graphs to generate molecular structures which could be synthesized could open new possibility for valid molecule generation and synthetic feasibility.

Neural Networks in the Design of Molecules with Affinity to Selected Protein Domains

The research proposed aims to create new chemical structures supported by a deep neural network that will possess an affinity to the selected protein domains, namely PDB IDs 7NPC, 7NP5, and 7KXD.

MEMES: Machine learning framework for Enhanced MolEcular Screening

A novel Machine learning framework for Enhanced MolEcular Screening (MEMES) based on Bayesian optimization is proposed for efficient sampling of the chemical space and is demonstrated to identify 90% of the top-1000 molecules from a molecular library of size about 100 million.

Retro Drug Design: From Target Properties to Molecular Structures

An AI strategy, retro drug design, or RDD, to generate novel small molecule drugs from scratch to meet predefined requirements, including but not limited to biological activity against a drug target, and optimal range of physicochemical and ADMET properties is described.

Seq2Mol: Automatic design of de novo molecules conditioned by the target protein sequences through deep neural networks.

A method which is conditioned on the protein target sequence to generate de novo molecules that are relevant to the target and generated compounds which are structurally different form the training set, while also being more similar to compounds known to bind to the two families of drug targets.

Structure-Based de Novo Molecular Generator Combined with Artificial Intelligence and Docking Simulations

A new deep learning-based molecular generator, SBMolGen, that integrates a recurrent neural network, a Monte Carlo tree search, and docking simulations that not only generates novel binding active molecules but also presents 3D docking poses with target proteins, which will be useful in subsequent drug design.

ChemistGA: A Chemical Synthesizable Accessible Molecular Generation Algorithm for Real-World Drug Discovery.

Calculations on the two benchmarks illustrate that ChemistGA achieves impressive performance among the state-of-the-art baselines, and it opens a new avenue for the application of generative models to real-world drug discovery scenarios.

De novo design of small molecules against drug targets of central nervous system using multi-property optimization

Results from the study show the capability of the proposed method to learn the molecular features required to produce novel small molecules with multiple desired physico-chemical properties against the target protein rapidly.

Pocket2Drug: An Encoder-Decoder Deep Neural Network for the Target-Based Drug Design

Pocket2Drug is a promising computational approach to inform the discovery of novel biopharmaceuticals that learns the conditional probability distribution of small molecules from a large dataset of pocket structures with supervised training, followed by the sampling of drug candidates from the trained model.
...

References

SHOWING 1-10 OF 75 REFERENCES

Chemical-Space-Based de Novo Design Method To Generate Drug-Like Molecules

The modified DAECS is modified to enable the user to select a target area to consider properties other than activity and improve the diversity of the generated structures by visualizing the drug-likeness distribution and the activity distribution.

Multi-objective molecular de novo design by adaptive fragment prioritization.

The nanomolar potencies of the hits obtained, their high ligand efficiencies, and an overall success rate of 90 % demonstrate that this ligand-based computer-aided molecular design method may guide target-focused combinatorial chemistry.

Development of a New De Novo Design Algorithm for Exploring Chemical Space

A new de novo design system to search a target area in chemical space and showed that the proposed generator could produce diverse virtual compounds that had high activity in docking simulations.

DOGS: Reaction-Driven de novo Design of Bioactive Compounds

A deterministic compound construction procedure that explicitly considers compound synthesizability is implemented, based on a compilation of 25'144 readily available synthetic building blocks and 58 established reaction principles, and proved to be suitable for scaffold-hopping from known ligands to novel chemotypes, and for generating bioactive molecules with drug-like properties.

Deep Architectures and Deep Learning in Chemoinformatics: The Prediction of Aqueous Solubility for Drug-Like Molecules

A brief overview of deep learning methods is presented and in particular how recursive neural network approaches can be applied to the problem of predicting molecular properties, by considering an ensemble of recursive neural networks associated with all possible vertex-centered acyclic orientations of the molecular graph.

Enabling future drug discovery by de novo design

An overview of the various methodologies for virtual compound construction, evaluation, and optimization in machina, and how they can support medicinal chemistry projects in the early phase of drug discovery is given.

Low Data Drug Discovery with One-Shot Learning

This work demonstrates how one-shot learning can be used to significantly lower the amounts of data required to make meaningful predictions in drug discovery applications and introduces a new architecture, the iterative refinement long short-term memory, that significantly improves learning of meaningful distance metrics over small-molecules.

Deep Learning in Drug Discovery

An overview of this emerging field of molecular informatics, the basic concepts of prominent deep learning methods are presented, and motivation to explore these techniques for their usefulness in computer‐assisted drug discovery and design is offered.

SCUBIDOO: A Large yet Screenable and Easily Searchable Database of Computationally Created Chemical Compounds Optimized toward High Likelihood of Synthetic Tractability

SCUBIDOO is a freely accessible database concept that currently holds 21 million virtual products originating from a small library of building blocks and a collection of robust organic reactions, and might be a useful idea generator for early ligand discovery projects since it allows a focus on those molecules that are likely to be synthetically feasible and can therefore be studied further.

The enumeration of chemical space

The chemical space of molecules following simple rules of chemical stability and synthetic feasibility has been enumerated up to 13 atoms of C, N, O, Cl, S, forming the GDB‐13 database with 977 million structures, which is organized in a 42‐dimensional chemical space using molecular quantum numbers (MQN) as descriptors.
...