scholarly journals An Exploration Strategy Improves the Diversity of de novo Ligands Using Deep Reinforcement Learning – A Case for the Adenosine A2A Receptor

2019 ◽  
Author(s):  
Xuhan Liu ◽  
Kai Ye ◽  
Herman Van Vlijmen ◽  
Adriaan P. IJzerman ◽  
Gerard JP Van westen
Keyword(s):  
Reinforcement Learning ◽  
Language Processing ◽  
De Novo ◽  
Chemical Space ◽  
Chemical Diversity ◽  
Reward Function ◽  
Chemical Structures ◽  
Exploration Strategy ◽  
Machine Learning Model ◽  
Adenosine A2a

<p></p><p>Over the last five years deep learning has progressed tremendously in both image recognition and natural language processing. Now it is increasingly applied to other data rich fields. In drug discovery, recurrent neural networks (RNNs) have been shown to be an effective method to generate novel chemical structures in the form of SMILES. However, ligands generated by current methods have so far provided relatively low diversity and do not fully cover the whole chemical space occupied by known ligands. Here, we propose a new method (DrugEx) to discover <i>de novo</i> drug-like molecules. DrugEx is an RNN model (generator) trained through reinforcement learning which was integrated with a special exploration strategy. As a case study we applied our method to design ligands against the adenosine A<sub>2A</sub> receptor. From ChEMBL data, a machine learning model (predictor) was created to predict whether generated molecules are active or not. Based on this predictor as the reward function, the generator was trained by reinforcement learning without any further data. We then compared the performance of our method with two previously published methods, REINVENT and ORGANIC. We found that candidate molecules our model designed, and predicted to be active, had a larger chemical diversity, and better covered the chemical space of known ligands compared to the state-of-the-art.</p><p></p>

scholarly journals An Exploration Strategy Improves the Diversity of de novo Ligands Using Deep Reinforcement Learning – A Case for the 3 Adenosine A2A Receptor

2018 ◽  
Author(s):  
Xuhan Liu ◽  
Kai Ye ◽  
Herman Van Vlijmen ◽  
Adriaan P. IJzerman ◽  
Gerard JP Van westen
Keyword(s):  
Reinforcement Learning ◽  
Language Processing ◽  
De Novo ◽  
Chemical Space ◽  
Chemical Diversity ◽  
Reward Function ◽  
Chemical Structures ◽  
Exploration Strategy ◽  
Machine Learning Model ◽  
Adenosine A2a

<p>Over the last five years deep learning has progressed tremendously in both image recognition and natural language processing. Now it is increasingly applied to other data rich fields. In drug discovery, recurrent neural networks (RNNs) have been shown to be an effective method to generate novel chemical structures in the form of SMILES. However, ligands generated by current methods used to provide relatively little diversity and do not fully cover the whole chemical space occupied by known ligands. Here, we propose a new method (DrugEx) to discover <i>de novo</i> drug-like molecules. DrugEx is an RNN model (generator) trained through a special exploration strategy integrated into reinforcement learning. As a case study we applied our method to design ligands against the adenosine A<sub>2A</sub> receptor. From ChEMBL data, a machine learning model (predictor) was created to predict whether generated molecules are active or not. Based on this predictor as the reward function, the generator was trained by reinforcement learning without any further data. We then compared the performance of our method with two previously published methods, REINVENT and ORGANIC. We found that candidate molecules our model designed that were predicted to be active, had a larger chemical diversity, and better covered the chemical space of known ligands compared to the state-of-the-art.</p>

scholarly journals An Exploration Strategy Improves the Diversity of de novo Ligands Using Deep Reinforcement Learning – A Case for the Adenosine A2A Receptor

2019 ◽  
Author(s):  
Xuhan Liu ◽  
Kai Ye ◽  
Herman Van Vlijmen ◽  
Adriaan P. IJzerman ◽  
Gerard JP Van westen
Keyword(s):  
Reinforcement Learning ◽  
Language Processing ◽  
De Novo ◽  
Chemical Space ◽  
Chemical Diversity ◽  
Reward Function ◽  
Chemical Structures ◽  
Exploration Strategy ◽  
Machine Learning Model ◽  
Adenosine A2a

<p></p><p>Over the last five years deep learning has progressed tremendously in both image recognition and natural language processing. Now it is increasingly applied to other data rich fields. In drug discovery, recurrent neural networks (RNNs) have been shown to be an effective method to generate novel chemical structures in the form of SMILES. However, ligands generated by current methods have so far provided relatively low diversity and do not fully cover the whole chemical space occupied by known ligands. Here, we propose a new method (DrugEx) to discover <i>de novo</i> drug-like molecules. DrugEx is an RNN model (generator) trained through reinforcement learning which was integrated with a special exploration strategy. As a case study we applied our method to design ligands against the adenosine A<sub>2A</sub> receptor. From ChEMBL data, a machine learning model (predictor) was created to predict whether generated molecules are active or not. Based on this predictor as the reward function, the generator was trained by reinforcement learning without any further data. We then compared the performance of our method with two previously published methods, REINVENT and ORGANIC. We found that candidate molecules our model designed, and predicted to be active, had a larger chemical diversity, and better covered the chemical space of known ligands compared to the state-of-the-art.</p><p></p>

scholarly journals An Exploration Strategy Improves the Diversity of de novo Ligands Using Deep Reinforcement Learning – A Case for the Adenosine A2A Receptor

2018 ◽  
Author(s):  
Xuhan Liu ◽  
Kai Ye ◽  
Herman Van Vlijmen ◽  
Adriaan P. IJzerman ◽  
Gerard JP Van westen
Keyword(s):  
Reinforcement Learning ◽  
Language Processing ◽  
De Novo ◽  
Chemical Space ◽  
Chemical Diversity ◽  
Reward Function ◽  
Chemical Structures ◽  
Exploration Strategy ◽  
Machine Learning Model ◽  
Adenosine A2a

<p>Over the last five years deep learning has progressed tremendously in both image recognition and natural language processing. Now it is increasingly applied to other data rich fields. In drug discovery, recurrent neural networks (RNNs) have been shown to be an effective method to generate novel chemical structures in the form of SMILES. However, ligands generated by current methods used to provide relatively little diversity and do not fully cover the whole chemical space occupied by known ligands. Here, we propose a new method (DrugEx) to discover <i>de novo</i> drug-like molecules. DrugEx is an RNN model (generator) trained through a special exploration strategy integrated into reinforcement learning. As a case study we applied our method to design ligands against the adenosine A<sub>2A</sub> receptor. From ChEMBL data, a machine learning model (predictor) was created to predict whether generated molecules are active or not. Based on this predictor as the reward function, the generator was trained by reinforcement learning without any further data. We then compared the performance of our method with two previously published methods, REINVENT and ORGANIC. We found that candidate molecules our model designed that were predicted to be active, had a larger chemical diversity, and better covered the chemical space of known ligands compared to the state-of-the-art.</p>

scholarly journals Memory-assisted reinforcement learning for diverse molecular de novo design

2020 ◽  
Author(s):  
Thomas Blaschke ◽  
Ola Engkvist ◽  
Jürgen Bajorath ◽  
Hongming Chen
Keyword(s):  
Reinforcement Learning ◽  
De Novo ◽  
Molecular Design ◽  
Chemical Space ◽  
Scoring Function ◽  
Chemical Diversity ◽  
Chemical Structures ◽  
Machine Learning Model ◽  
Low Diversity ◽  
De Novo Molecular Design

Abstract In de novo molecular design, recurrent neural networks (RNN) have been shown to be effective methods for sampling and generating novel chemical structures. Using a technique called reinforcement learning (RL), an RNN can be tuned to target a particular section of chemical space with optimized desirable properties using a scoring function. However, ligands generated by current RL methods so far tend to have relatively low diversity, and sometimes even result in duplicate structures when optimizing towards desired properties. Here, we propose a new method to address the low diversity issue in RL for molecular design. Memory-assisted RL is an extension of the known RL, with the introduction of a so-called memory unit. As proof of concept, we applied our method to generate structures with a desired AlogP value. In a second case study, we applied our method to design ligands for the dopamine type 2 receptor and the 5-hydroxytryptamine type 1A receptor. For both receptors, a machine learning model was developed to predict whether generated molecules were active or not for the receptor. In both case studies, it was found that memory-assisted RL led to the generation of more compounds predicted to be active having higher chemical diversity, thus achieving better coverage of chemical space of known ligands compared to established RL methods.

scholarly journals Memory-assisted reinforcement learning for diverse molecular de novo design

2020 ◽  
Vol 12 (1) ◽  
Author(s):  
Thomas Blaschke ◽  
Ola Engkvist ◽  
Jürgen Bajorath ◽  
Hongming Chen
Keyword(s):  
Reinforcement Learning ◽  
De Novo ◽  
Molecular Design ◽  
Chemical Space ◽  
Scoring Function ◽  
Chemical Diversity ◽  
Chemical Structures ◽  
Machine Learning Model ◽  
Low Diversity ◽  
De Novo Molecular Design

Abstract In de novo molecular design, recurrent neural networks (RNN) have been shown to be effective methods for sampling and generating novel chemical structures. Using a technique called reinforcement learning (RL), an RNN can be tuned to target a particular section of chemical space with optimized desirable properties using a scoring function. However, ligands generated by current RL methods so far tend to have relatively low diversity, and sometimes even result in duplicate structures when optimizing towards desired properties. Here, we propose a new method to address the low diversity issue in RL for molecular design. Memory-assisted RL is an extension of the known RL, with the introduction of a so-called memory unit. As proof of concept, we applied our method to generate structures with a desired AlogP value. In a second case study, we applied our method to design ligands for the dopamine type 2 receptor and the 5-hydroxytryptamine type 1A receptor. For both receptors, a machine learning model was developed to predict whether generated molecules were active or not for the receptor. In both case studies, it was found that memory-assisted RL led to the generation of more compounds predicted to be active having higher chemical diversity, thus achieving better coverage of chemical space of known ligands compared to established RL methods.

scholarly journals Memory-assisted Reinforcement Learning for Diverse Molecular De Novo Design

2020 ◽  
Author(s):  
Thomas Blaschke ◽  
Ola Engkvist ◽  
Jürgen Bajorath ◽  
Hongming Chen
Keyword(s):  
Reinforcement Learning ◽  
De Novo ◽  
Molecular Design ◽  
Chemical Space ◽  
Scoring Function ◽  
Chemical Diversity ◽  
Chemical Structures ◽  
Low Diversity ◽  
De Novo Molecular Design ◽  
Dopamine 2 Receptor

Abstract In de novo molecular design, recurrent neural networks (RNN) have been shown to be effective methods for sampling and generating novel chemical structures. Using a technique called reinforcement learning (RL), an RNN can be tuned to target a particular section of chemical space with optimized desirable properties using a scoring function. However, ligands generated by current RL methods so far tend to have relatively low diversity, and sometimes even result in duplicate structures when optimizing towards particular properties. Here, we propose a new method to address the low diversity issue in RL. Memory-assisted RL is an extension of the known RL, with the introduction of a so-called memory unit. As proof of concept, we applied our method to generate structures with an optimized logP. In a second case study, we applied our method to design ligands for the dopamine 2 receptor and the 5-hydroxytryptamine 1A receptor. For both receptors, a machine learning model was developed to predict whether generated molecules were active or not for the receptor. In both case studies, it was found that memory-assisted RL led to the generation of more active compounds and with higher chemical diversity, thus achieving better coverage of chemical space of known ligands compared to established RL method.

scholarly journals Memory-Assisted Reinforcement Learning for Diverse Molecular De Novo Design

2020 ◽  
Author(s):  
Thomas Blaschke ◽  
Ola Engkvist ◽  
Jürgen Bajorath ◽  
Hongming Chen
Keyword(s):  
Reinforcement Learning ◽  
De Novo ◽  
Molecular Design ◽  
Chemical Space ◽  
Scoring Function ◽  
De Novo Design ◽  
Memory Unit ◽  
Chemical Structures ◽  
Low Diversity ◽  
De Novo Molecular Design

<div><div><div><p>In de novo molecular design, recurrent neural networks (RNN) have been shown to be effective methods for sampling and generating novel chemical structures. Using a technique called reinforcement learning (RL), an RNN can be tuned to target a particular section of chemical space with optimized desirable properties using a scoring function. However, ligands generated by current RL methods so far tend to have relatively low diversity, and sometimes even result in duplicate structures when optimizing towards particular properties. Here, we propose a new method to address the low diversity issue in RL. Memory-assisted RL is an extension of the known RL, with the introduction of a so-called memory unit.</p></div></div></div>

scholarly journals Memory-Assisted Reinforcement Learning for Diverse Molecular De Novo Design

2020 ◽  
Author(s):  
Thomas Blaschke ◽  
Ola Engkvist ◽  
Jürgen Bajorath ◽  
Hongming Chen
Keyword(s):  
Reinforcement Learning ◽  
De Novo ◽  
Molecular Design ◽  
Chemical Space ◽  
Scoring Function ◽  
De Novo Design ◽  
Memory Unit ◽  
Chemical Structures ◽  
Low Diversity ◽  
De Novo Molecular Design

<div><div><div><p>In de novo molecular design, recurrent neural networks (RNN) have been shown to be effective methods for sampling and generating novel chemical structures. Using a technique called reinforcement learning (RL), an RNN can be tuned to target a particular section of chemical space with optimized desirable properties using a scoring function. However, ligands generated by current RL methods so far tend to have relatively low diversity, and sometimes even result in duplicate structures when optimizing towards particular properties. Here, we propose a new method to address the low diversity issue in RL. Memory-assisted RL is an extension of the known RL, with the introduction of a so-called memory unit.</p></div></div></div>

scholarly journals Analysis of a large food chemical database: chemical space, diversity, and complexity

F1000Research ◽  
2018 ◽  
Vol 7 ◽  
pp. 993 ◽  
Author(s):  
J. Jesús Naveja ◽  
Mariel P. Rico-Hidalgo ◽  
José L. Medina-Franco
Keyword(s):  
Natural Products ◽  
Chemical Space ◽  
Structural Complexity ◽  
Chemical Diversity ◽  
Chemical Structures ◽  
Novel Approach ◽  
Chemical Database ◽  
Protein Interactome ◽  
Limited Basis ◽  
Future Direction

Background: Food chemicals are a cornerstone in the food industry. However, its chemical diversity has been explored on a limited basis, for instance, previous analysis of food-related databases were done up to 2,200 molecules. The goal of this work was to quantify the chemical diversity of chemical compounds stored in FooDB, a database with nearly 24,000 food chemicals. Methods: The visual representation of the chemical space of FooDB was done with ChemMaps, a novel approach based on the concept of chemical satellites. The large food chemical database was profiled based on physicochemical properties, molecular complexity and scaffold content. The global diversity of FooDB was characterized using Consensus Diversity Plots. Results: It was found that compounds in FooDB are very diverse in terms of properties and structure, with a large structural complexity. It was also found that one third of the food chemicals are acyclic molecules and ring-containing molecules are mostly monocyclic, with several scaffolds common to natural products in other databases. Conclusions: To the best of our knowledge, this is the first analysis of the chemical diversity and complexity of FooDB. This study represents a step further to the emerging field of “Food Informatics”. Future study should compare directly the chemical structures of the molecules in FooDB with other compound databases, for instance, drug-like databases and natural products collections. An additional future direction of this work is to use the list of 3,228 polyphenolic compounds identified in this work to enhance the on-going polyphenol-protein interactome studies.

Sign in / Sign up

Export Citation Format

Share Document