Deep Learning in Drug Design : Methods and Applications / Qifeng Bai, Tingyang Xu, Junzhou Huang.

Format:

Book

Contributor:

ScienceDirect (Online service)

Language:

English

Subjects (All):

Drugs--Design--Data processing.

Drugs.

Drug Design--methods.

Medical Subjects:

Drug Design--methods.

Physical Description:

1 online resource (xxv, 472 pages)

Place of Publication:

London ; Cambridge, MA : Academic Press, an imprint of Elsevier, [2026]

Contents:

Front Cover

Deep Learning in Drug Design

Copyright

Contents

List of contributors

Preface

Acknowledgments

1 Deep learning theories and methods for drug design

1 Molecular representations in deep learning

1.1 Introduction

1.2 One-dimensional (1D) molecular representations and applications

1.3 Two-dimensional (2D) molecular representations and applications

1.4 Three-dimensional (3D) molecular representations and applications

1.5 Molecular multimodality

1.6 Conclusion

References

2 CNNs in drug design

2.1 Introduction

2.2 Preliminaries

2.2.1 Convolutional neural networks

2.2.2 CNN for biological data

2.3 Drug-target interactions prediction

2.3.1 Drug-based models

2.3.2 Structure-based models

2.3.3 Drug-protein-based models

2.4 Drug sensitivity and response prediction

2.5 Drug-drug interactions side effect prediction

2.6 Drug-drug similarity prediction

2.7 Conclusion

3 GNNs in drug design

3.1 Introduction

3.1.1 Drug design

3.1.2 Computer-aided drug design

3.1.3 Deep learning in drug design

3.1.4 GNNs in drug design

3.1.5 SOTA structure-based models

3.2 Small molecule drug design

3.2.1 Virtual screening

3.2.2 De novo design

3.2.3 Dataset

3.3 Antibody design

3.3.1 Antibody

3.3.2 Inverse folding task

3.3.3 Antibody design based on inverse folding task

3.3.4 Dataset

3.4 Peptide therapeutics

3.4.1 Dataset

3.5 Conclusion and discussion

4 RNNs and LSTM in drug design

4.1 Introduction

4.2 Theories and classification of RNNs

4.2.1 Long short-term memory

4.2.2 Bidirectional recurrent neural networks

4.2.3 Gated recurrent units

4.2.4 Other networks

4.3 Application of different network architectures of RNN in drug design

4.3.1 Simple RNNs

4.3.2 Encoder-decoder.

4.3.3 Attention RNNs

4.3.4 Reinforcement learning-based RNNs

4.3.5 Transfer learning-based RNNs

4.4 Conclusion and future prospects

5 Deep reinforcement learning in drug design

5.1 Introduction

5.2 Deep reinforcement learning theory

5.2.1 Discriminative and generative models

5.2.2 Deep reinforcement learning framework

5.2.3 Mainstream deep reinforcement learning methods and their application in drug design

5.3 Value-based deep reinforcement learning methods for drug design

5.3.1 Value-based deep reinforcement learning methods

5.3.2 Q-Learning algorithm for drug design

5.3.3 Deep Q-Network (DQN) algorithm for drug design

5.4 Policy-based deep reinforcement learning approach for drug design

5.4.1 Policy-based deep reinforcement learning methods

5.4.2 Policy gradient algorithms for drug design

5.4.3 PPO algorithm for drug design

5.5 Deep reinforcement learning methods based on actor-critic approaches for drug design

5.5.1 Deep reinforcement learning methods based on actor-critic approaches

5.5.2 Actor-critic algorithms for drug design

5.5.3 Deep Deterministic Policy Gradient (DDPG) for drug design

5.6 Summary and outlook

6 Transformer and drug design

6.1 Introduction

6.2 Preliminaries

6.3 Drug understanding and generation

6.3.1 Transformers for textual sequence

6.3.2 Graph Transformers for molecular structure

6.3.2.1 Graph Transformers for 2D molecules

6.3.2.2 Graph Transformers for 3D molecules

6.3.3 Protein design

6.3.4 Protein-molecule binding

6.4 Conclusion

7 Generative models for drug design

7.1 Introduction

7.2 Deep generative models

7.2.1 Generative adversarial networks

7.2.2 Variational autoencoders

7.2.3 Flow-based models

7.2.4 Diffusion models

7.2.5 Transformers.

7.2.6 Techniques for model enhancement

7.3 Overview of small molecule design

7.3.1 Task classification

7.3.1.1 Target-agnostic molecule design

7.3.1.2 Target-aware molecule design

7.3.2 Strategies

7.3.2.1 Ways for building blocks

7.3.2.2 Property restraint

7.3.2.3 3D molecule generation

7.3.2.4 Molecule generation for target protein

7.3.3 Evaluation metrics

7.3.4 Datasets

7.4 Overview of protein generation

7.4.1 Sequence-to-structure

7.4.2 Properties-to-sequence/structure

7.4.3 Struction-to-sequence

7.4.4 Backbone design

7.4.5 Sequence and structure co-design

7.5 Conclusions and future research needs

8 Geometric graph learning for drug design

8.1 Introduction

8.2 Model: geometric GNNs

8.2.1 Message-passing neural networks

8.2.2 Invariant graph neural networks

8.2.3 Equivariant graph neural networks

8.2.3.1 Scalarization-based models

8.2.3.2 Tensor-product-based models

8.2.4 Spherical-scalarization models

8.2.5 Geometric graph transformers

8.2.6 Theoretical analysis on expressivity

8.3 Conclusion

9 Self-supervised learning for drug discovery

9.1 Introduction

9.2 Self-supervised learning method on molecular representation

9.2.1 Generative learning on molecular representation

9.2.2 Contrastive learning on molecular representation

9.3 Self-supervised learning method on protein representation

9.3.1 Generative learning on protein representation

9.3.2 Contrastive learning on protein representation

9.4 Self-supervised learning method in drug-target binding affinity

9.5 Conclusion and discussion

10 Transfer learning and meta-learning for drug discovery

10.1 Introduction

10.2 Transfer learning

10.2.1 Categorization of transfer learning

10.2.2 Deep transfer learning.

10.2.2.1 Fine-tuning

10.2.2.2 Feature-based deep transfer learning

10.2.3 Transfer learning for drug discovery

10.3 Meta-learning

10.3.1 Typical meta-learning algorithms

10.3.1.1 Metric-based methods

10.3.1.2 Gradient-based methods

10.3.2 Meta-learning for drug discovery

10.4 Multi-task learning

10.4.1 Network-sharing architectures

10.4.1.1 Hard parameter sharing

10.4.1.2 Soft parameter sharing

10.4.2 Joint optimization strategies

10.4.2.1 Multi-task learning for drug discovery

10.5 Conclusion

11 Explainable artificial intelligence for drug design models

11.1 Introduction

11.2 Explainable artificial intelligence methods

11.2.1 The role of interpretability in AI

11.2.2 XAI modeling framework

11.2.3 Main XAI methods

11.3 Intrinsic interpretability models in drug discovery and design

11.3.1 Applications of decision trees in drug discovery and design

11.3.2 Applications of KNN in drug discovery and design

11.4 The application of post-hoc interpretable artificial intelligence in drug design

11.4.1 Application of SHAP in drug design

11.5 XAI methods for DNN in drug discovery and design

11.5.1 Adversarial explanation methods

11.5.2 Gradient-based explanation methods

11.6 Conclusion and prospects

12 Large models in drug design

12.1 Introduction

12.2 Basic principles of large models

12.2.1 Multi-layer perceptron and Kolmogorov-Arnold networks

12.2.2 Transformer

12.3 Methods of large models

12.3.1 Prompting methods

12.3.2 Pre-training and fine-tuning methods

12.3.3 Retrieval-augmented generation

12.3.4 Web search

12.4 Applications of large models in drug design

12.4.1 Applications of LLMs in small molecule research

12.4.2 Applications of LLMs in protein research

12.4.3 Applications of LLMs in gene research.

12.5 Conclusions

2 Deep learning applications in drug design

13 Deep learning for protein secondary structure prediction

13.1 Introduction

13.2 General pipeline description

13.3 Data

13.4 Position-specific scoring matrix

13.4.1 Multiple sequence alignment

13.4.2 PSSMs calculation

13.5 Secondary structure prediction models

13.5.1 Model architectures

13.5.2 Study on the labels

13.5.3 Input feature enhancement

13.6 Post-AlphaFold discussion

13.6.1 The impact of AlphaFold on structural bioinformatics

13.6.2 Future directions and post-AlphaFold secondary structure study

13.7 Conclusion

14 Deep learning in protein structure prediction

14.1 Introduction

14.2 Preliminaries

14.3 The evolution of protein structure prediction

14.3.1 Traditional methods

14.3.2 Deep learning-assisted methods

14.3.3 End-to-end deep learning methods

14.3.4 Deep learning methods for auxiliary tasks

14.4 Datasets and evaluation

14.5 Challenges and future directions

15 Deep learning for affinity prediction and interface prediction in molecular interactions

15.1 Introduction

15.2 Affinity prediction

15.2.1 Drug-target affinity prediction

15.2.2 Protein-ligand affinity prediction

15.2.3 Protein-protein affinity prediction

15.2.4 Others

15.3 Interface prediction

15.3.1 Protein-protein interaction interface prediction

15.3.2 Protein-ligand binding site prediction

15.4 Conclusion

16 Deep learning for complex structure prediction in molecular interactions

16.1 Introduction

16.2 Protein-protein complex structure prediction

16.3 Protein-ligand complex structure prediction

16.4 Generalist methods

16.5 Conclusion

17 Deep learning in chemical synthesis and retrosynthesis.

Notes:

Includes bibliographical references and index.

Electronic reproduction. Amsterdam Available via World Wide Web.

Description based on online resource; title from digital title page (viewed on May 14, 2026).

ISBN:

9780443329098

0443329095

Publisher Number:

90104286214

Access Restriction:

Restricted for use by site license.