Deep Learning in Protein Structural Modeling and Design

  title={Deep Learning in Protein Structural Modeling and Design},
  author={Wenhao Gao and Sai Pooja Mahajan and Jeremias Sulam and Jeffrey J. Gray},

Figures from this paper

Structure-based protein design with deep learning.
Deep Learning-Based Advances in Protein Structure Prediction
Important milestones and progresses in the field of protein structure prediction due to DL-based methods as observed in CASP experiments are highlighted and advances in various steps ofprotein structure prediction pipeline viz. protein contact map prediction, protein distogram prediction,protein real-valued distance prediction, and Quality Assessment/refinement are described.
Protein Design with Deep Learning
The representations used so far are described, their strengths and weaknesses are discussed, and their associated DL architecture for design and related tasks are details.
Protein loop modeling and refinement using deep learning models
Two novel deep learning architectures for loop modeling are proposed: one uses a combined convolutional neural network-recursive neural network (RNN) structure and the other is based on refinement of histograms using a 2D CNN architecture (DeepHisto).
Current directions in combining simulation-based macromolecular modeling approaches with deep learning
  • V. Mulligan
  • Computer Science
    Expert opinion on drug discovery
  • 2021
This review introduces biologists to deep neural network architecture, surveys recent successes of deep learning in structure prediction, and discusses emerging deep learning-based approaches for structure-function analysis and design, with particular focus on the interplay between simulation-based and neural network- based approaches.
Deep Generative Modeling for Protein Design
100th Anniversary of Macromolecular Science Viewpoint: Data-Driven Protein Design.
This Viewpoint reviews the applications of machine learning and, in particular, deep learning as part of data-driven protein engineering platforms and highlights recent successes, review promising computational methodologies, and provides an outlook on future challenges and opportunities.
Antibody structure prediction using interpretable deep learning
DeepAb, a deep learning method for predicting accurate antibody FV structures from sequence, is presented and improved accuracy is demonstrated and interpretable outputs about specific amino acids and residue interactions are revealed that should facilitate design of novel therapeutic antibodies.
How Deep Learning Tools Can Help Protein Engineers Find Good Sequences.
The use of deep learning tools to find good sequences for protein engineering are reviewed, including developing oracles/predictors of a property of the proteins and methods that sample from a distribution of protein-like sequences to optimize the desired property.
Deep learning of Protein Sequence Design of Protein-protein Interactions
An attention-based deep learning model inspired by algorithms used for image-caption assignments for sequence design of peptides or protein fragments is developed which allows the one-sided design of a given protein fragment which can be applicable for the redesign of protein-interfaces or the de novo design of new interactions fragments.


Deep learning methods in protein structure prediction
Accurate Protein Structure Prediction by Embeddings and Deep Learning Representations
This work demonstrates state-of-the-art protein structure prediction (PSP) results using embeddings and deep learning models for prediction of backbone atom distance matrices and torsion angles, and creates a new gold standard dataset of proteins which is comprehensive and easy to use.
Unified rational protein engineering with sequence-based deep representation learning
Deep learning is applied to unlabeled amino-acid sequences to distill the fundamental features of a protein into a statistical representation that is semantically rich and structurally, evolutionarily and biophysically grounded and broadly applicable to unseen regions of sequence space.
Generative Models for Graph-Based Protein Design
This framework significantly improves in both speed and robustness over conventional and deep-learning-based methods for structure-based protein sequence design, and takes a step toward rapid and targeted biomolecular design with the aid of deep generative models.
Generative modeling for protein structures
This work applies Generative Adversarial Networks (GANs) to the task of generating protein structures, and encodes protein structures in terms of pairwise distances between alpha-carbons on the protein backbone, which eliminates the need for the generative model to learn translational and rotational symmetries.
Recent developments in deep learning applied to protein structure prediction
A brief introduction to some of the key principles and properties of DNN models is offered and why they are naturally suited to certain problems in structural bioinformatics is discussed, as well as a discussion on potential pitfalls.
Protein sequence design with a learned potential
This work investigates the capability of a deep neural network model to automate design of sequences onto protein backbones, having learned directly from crystal structure data and without any human-specified priors.
Computational Protein Design with Deep Learning Neural Networks
Using the network output as residue type restraints improves the average sequence identity in designing three natural proteins using Rosetta, and the predictions from the network show ~3% higher sequence identity than a previous method.
Finding the needle in the haystack: towards solving the protein-folding problem computationally
This review aims at summarizing conformational sampling algorithms and energy approximations that have been frequently used to study protein-folding mechanisms or to de novo predict protein tertiary structures and concludes with a summary of current challenges faced and an outlook on potential future directions.
Solving the protein folding problem in hydrophobic-polar model using deep reinforcement learning
A comparison between the DRL approach and other notable approaches shows that the approach outperforms other approaches in finding the minimum value of the free energy, which is the main factor in the protein folding problem, in less time in any available case.