Significant research on deep neural networks, culminating in AlphaFold2, convincingly shows that deep learning can predict the na- tive structure of a given protein sequence with high accuracy. In contrast, work on deep learning frameworks that can account for the structural plasticity of protein molecules remains in its infancy. Many researchers are now investigating deep generative models to explore the structure space of a protein. Current models largely use 2D convolution, leveraging representations of protein structures as contact maps or distance matri- ces. The goal is exclusively to generate protein-like, sequence-agnostic tertiary structures, but no rigorous metrics are utilized to convincingly make this case. This paper makes several contributions. It builds on momentum in graph representation learning and formalizes a protein tertiary structure as a contact graph. It demonstrates that graph repre- sentation learning outperforms models based on image convolution. This work also equips graph-based deep latent variable models with the abil- ity to learn from experimentally-available tertiary structures of proteins of varying lengths. The resulting models are shown to outperform state- of-the-art ones on rigorous metrics that quantify both local and distal patterns in physically-realistic protein structures. We hope this work will spur further research in deep generative models for obtaining a broader view of the structure space of a protein molecule.
more »
« less
Generating tertiary protein structures via interpretable graph variational autoencoders
Modeling the structural plasticity of protein molecules remains challenging. Most research has focused on obtaining one biologically active structure. This includes the recent AlphaFold2 that has been hailed as a breakthrough for protein modeling. Computing one structure does not suffice to understand how proteins modulate their interactions and even evade our immune system. Revealing the structure space available to a protein remains challenging. Data-driven approaches that learn to generate tertiary structures are increasingly garnering attention. These approaches exploit the ability to represent tertiary structures as contact or distance maps and make direct analogies with images to harness convolution-based generative adversarial frameworks from computer vision. Since such opportunistic analogies do not allow capturing highly structured data, current deep models struggle to generate physically realistic tertiary structures.
We present novel deep generative models that build upon the graph variational autoencoder framework. In contrast to existing literature, we represent tertiary structures as ‘contact’ graphs, which allow us to leverage graph-generative deep learning. Our models are able to capture rich, local and distal constraints and additionally compute disentangled latent representations that reveal the impact of individual latent factors. This elucidates what the factors control and makes our models more interpretable. Rigorous comparative evaluation along various metrics shows that the models, we propose advance the state-of-the-art. While there is still much ground to cover, the work presented here is an important first step, and graph-generative frameworks promise to get us to our goal of unraveling the exquisite structural complexity of protein molecules.
Code is available at https://github.com/anonymous1025/CO-VAE.
Supplementary data are available at Bioinformatics Advances online.
- NSF-PAR ID:
- 10361146
- Publisher / Repository:
- Oxford University Press
- Date Published:
- Journal Name:
- Bioinformatics Advances
- Volume:
- 1
- Issue:
- 1
- ISSN:
- 2635-0041
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
more » « less
Abstract Motivation Expanding our knowledge of small molecules beyond what is known in nature or designed in wet laboratories promises to significantly advance cheminformatics, drug discovery, biotechnology and material science. In silico molecular design remains challenging, primarily due to the complexity of the chemical space and the non-trivial relationship between chemical structures and biological properties. Deep generative models that learn directly from data are intriguing, but they have yet to demonstrate interpretability in the learned representation, so we can learn more about the relationship between the chemical and biological space. In this article, we advance research on disentangled representation learning for small molecule generation. We build on recent work by us and others on deep graph generative frameworks, which capture atomic interactions via a graph-based representation of a small molecule. The methodological novelty is how we leverage the concept of disentanglement in the graph variational autoencoder framework both to generate biologically relevant small molecules and to enhance model interpretability.
Results Extensive qualitative and quantitative experimental evaluation in comparison with state-of-the-art models demonstrate the superiority of our disentanglement framework. We believe this work is an important step to address key challenges in small molecule generation with deep generative frameworks.
Availability and implementation Training and generated data are made available at https://ieee-dataport.org/documents/dataset-disentangled-representation-learning-interpretable-molecule-generation. All code is made available at https://anonymous.4open.science/r/D-MolVAE-2799/.
Supplementary information Supplementary data are available at Bioinformatics online.
-
more » « less
Abstract Motivation Quality assessment (QA) of predicted protein tertiary structure models plays an important role in ranking and using them. With the recent development of deep learning end-to-end protein structure prediction techniques for generating highly confident tertiary structures for most proteins, it is important to explore corresponding QA strategies to evaluate and select the structural models predicted by them since these models have better quality and different properties than the models predicted by traditional tertiary structure prediction methods.
Results We develop EnQA, a novel graph-based 3D-equivariant neural network method that is equivariant to rotation and translation of 3D objects to estimate the accuracy of protein structural models by leveraging the structural features acquired from the state-of-the-art tertiary structure prediction method—AlphaFold2. We train and test the method on both traditional model datasets (e.g. the datasets of the Critical Assessment of Techniques for Protein Structure Prediction) and a new dataset of high-quality structural models predicted only by AlphaFold2 for the proteins whose experimental structures were released recently. Our approach achieves state-of-the-art performance on protein structural models predicted by both traditional protein structure prediction methods and the latest end-to-end deep learning method—AlphaFold2. It performs even better than the model QA scores provided by AlphaFold2 itself. The results illustrate that the 3D-equivariant graph neural network is a promising approach to the evaluation of protein structural models. Integrating AlphaFold2 features with other complementary sequence and structural features is important for improving protein model QA.
Availability and implementation The source code is available at https://github.com/BioinfoMachineLearning/EnQA.
Supplementary information Supplementary data are available at Bioinformatics online.
-
more » « less
Abstract Motivation Deep learning has revolutionized protein tertiary structure prediction recently. The cutting-edge deep learning methods such as AlphaFold can predict high-accuracy tertiary structures for most individual protein chains. However, the accuracy of predicting quaternary structures of protein complexes consisting of multiple chains is still relatively low due to lack of advanced deep learning methods in the field. Because interchain residue–residue contacts can be used as distance restraints to guide quaternary structure modeling, here we develop a deep dilated convolutional residual network method (DRCon) to predict interchain residue–residue contacts in homodimers from residue–residue co-evolutionary signals derived from multiple sequence alignments of monomers, intrachain residue–residue contacts of monomers extracted from true/predicted tertiary structures or predicted by deep learning, and other sequence and structural features.
Results Tested on three homodimer test datasets (Homo_std dataset, DeepHomo dataset and CASP-CAPRI dataset), the precision of DRCon for top L/5 interchain contact predictions (L: length of monomer in a homodimer) is 43.46%, 47.10% and 33.50% respectively at 6 Å contact threshold, which is substantially better than DeepHomo and DNCON2_inter and similar to Glinter. Moreover, our experiments demonstrate that using predicted tertiary structure or intrachain contacts of monomers in the unbound state as input, DRCon still performs well, even though its accuracy is lower than using true tertiary structures in the bound state are used as input. Finally, our case study shows that good interchain contact predictions can be used to build high-accuracy quaternary structure models of homodimers.
Availability and implementation The source code of DRCon is available at https://github.com/jianlin-cheng/DRCon. The datasets are available at https://zenodo.org/record/5998532#.YgF70vXMKsB.
Supplementary information Supplementary data are available at Bioinformatics online.
-
null (Ed.)Protein molecules are inherently dynamic and modulate their interactions with different molecular partners by accessing different tertiary structures under physiological conditions. Elucidating such structures remains challenging. Current momentum in deep learning and the powerful performance of generative adversarial networks (GANs) in complex domains, such as computer vision, inspires us to investigate GANs on their ability to generate physically-realistic protein tertiary structures. The analysis presented here shows that several GAN models fail to capture complex, distal structural patterns present in protein tertiary structures. The study additionally reveals that mechanisms touted as effective in stabilizing the training of a GAN model are not all effective, and that performance based on loss alone may be orthogonal to performance based on the quality of generated datasets. A novel contribution in this study is the demonstration that Wasserstein GAN strikes a good balance and manages to capture both local and distal patterns, thus presenting a first step towards more powerful deep generative models for exploring a possibly very diverse set of structures supporting diverse activities of a protein molecule in the cell.more » « less
