Orientation-Aware Graph Neural Networks for Protein Structure Representation Learning
Peking University（北京大学）；Tsinghua University（清华大学）；Massachusetts Institute of Technology（麻省理工学院）；Stanford University（斯坦福大学）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Helixon Ltd.（Helixon有限公司）
By folding into particular 3D structures, proteins play a key role in living beings. To learn meaningful representation from a protein structure for downstream tasks, not only the global backbone topology but the local fine-grained orientational relations between amino acids should also be considered. In this work, we propose the Orientation-Aware Graph Neural Networks (OAGNNs) to better sense the geometric characteristics in protein structure (e.g. inner-residue torsion angles, inter-residue orientations). Extending a single weight from a scalar to a 3D vector, we construct a rich set of geometric-meaningful operations to process both the classical and SO(3) representations of a given structure. To plug our designed perceptron unit into existing Graph Neural Networks, we further introduce an equivariant message passing paradigm, showing superior versatility in maintaining SO(3)-equivariance at the global scale. Experiments have shown that our OAGNNs have a remarkable ability to sense geometric orientational features compared to classical networks. OAGNNs have also achieved state-of-the-art performance on various computational biology applications related to protein 3D structures. The code is available at https://github.com/Ced3-han/OAGNN/tree/main.
GLProtein: Global-and-Local Structure Aware Protein Representation Learning
Department of Computing (COMP), The Hong Kong Polytechnic University（计算机学院（COMP），香港理工大学）；Department of Management and Marketing (MM), The Hong Kong Polytechnic University（管理与营销学院（MM），香港理工大学）；Department of Computer Science, Sichuan University（计算机学院，四川大学）
Rigidity-Aware Geometric Pretraining for Protein Design and Conformational Ensembles
University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；University of Washington（华盛顿大学）；MPI for Intelligent Systems, Tübingen（智能系统马克斯·普朗克研究所，图宾根）；National Research Council of Canada（加拿大国家研究理事会）；University of Ottawa（渥太华大学）；The Chinese University of Hong Kong（香港中文大学）
Protein Secondary Structure Prediction Using 3D Graphs and Relation-Aware Message Passing Transformers
Banaras Hindu University, Varanasi, India（班纳拉锡大学）；Indiana University, Bloomington（印第安纳大学）；Indian Institute of Technology Jodhpur, India（印度理工学院朱罗尔分校）；Indian Institute of Technology Patna, India（印度理工学院帕坦分校）
Structure-Aware Compound-Protein Affinity Prediction via Graph Neural Networks with Group Lasso Regularization
Department of Biostatistics & Health Data Science（生物统计学与健康数据科学系）；Indiana University（印第安纳大学）；Department of Computer Science（计算机科学系）；Indiana University Bloomington（印第安纳大学布卢明顿分校）；Division of Clinical Pharmacology（临床药理学部）；Indiana University School of Medicine（印第安纳大学医学院）；IUSM-Purdue TREAT-AD Center（IUSM-普渡大学TREAT-AD中心）；Department of Medical and Molecular Genetics（医学与分子遗传学系）
University of Science and Technology of China（中国科学技术大学）；Baidu Research（百度研究院）；Baidu Inc.（百度公司）；Baidu Research USA（百度美国研究院）；Oregon State University（俄勒冈州立大学）；HWL Consulting LLC（HWL咨询有限责任公司）；Rutgers University（罗格斯大学）
Drug discovery often relies on the successful prediction of protein-ligand binding affinity. Recent advances have shown great promise in applying graph neural networks (GNNs) for better affinity prediction by learning the representations of protein-ligand complexes. However, existing solutions usually treat protein-ligand complexes as topological graph data, thus the biomolecular structural information is not fully utilized. The essential long-range interactions among atoms are also neglected in GNN models. To this end, we propose a structure-aware interactive graph neural network (SIGN) which…
EGRL: Edge generation-guided relation-aware learning for RNA-protein interaction prediction
Macau University of Science and Technology（澳门科技大学）；Macau University of Science and Technology Zhuhai Research Institute（澳门科技大学珠海研究院）；Wuhan University of Technology（武汉理工大学）；Jiangsu University（江苏大学）；Harbin Institute of Technology（哈尔滨工业大学）
RIGA-Fold: A General Framework for Protein Inverse Folding via Recurrent Interaction and Geometric Awareness
College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, China（深圳大学计算机科学与软件工程学院）；National Engineering Laboratory for Big Data System Computing Technology, Shenzhen University, Shenzhen, China（大数据系统计算技术国家工程实验室）
AI Research and Innovation Lab, Westlake University（西湖大学人工智能研究与创新实验室）；Zhejiang University（浙江大学）
While DeepMind has tentatively solved protein folding, its inverse problem -- protein design which predicts protein sequences from their 3D structures -- still faces significant challenges. Particularly, the lack of large-scale standardized benchmark and poor accuray hinder the research progress. In order to standardize comparisons and draw more research interest, we use AlphaFold DB, one of the world's largest protein structure databases, to establish a new graph-based benchmark -- AlphaDesign. Based on AlphaDesign, we propose a new method called ADesign to improve accuracy by introducing pro…
Rotation-Adaptive Point Cloud Domain Generalization via Intricate Orientation Learning
South China University of Technology（华南理工大学）；Guangdong Engineering Center for Large Model and GenAI Technology（广东省大模型与生成式人工智能技术工程中心）；State Key Laboratory of Subtropical Building and Urban Science（亚热带建筑与城市科学国家重点实验室）；Ministry of Education Key Laboratory of Big Data and Intelligent Robot（大数据与智能机器人教育部重点实验室）；Singapore Management University（新加坡管理大学）
Lawrence Livermore National Laboratory（劳伦斯利弗莫尔国家实验室）
Improved understanding of the relation between the behavior of RAS and RAF proteins and the local lipid environment in the cell membrane is critical for getting insights into the mechanisms underlying cancer formation. In this work, we employ deep learning (DL) to learn this relationship by predicting protein orientational states of RAS and RAS-RAF protein complexes with respect to the lipid membrane based on the lipid densities around the protein domains from coarse-grained (CG) molecular dynamics (MD) simulations. Our DL model can predict six protein states with an overall accuracy of over 8…
A method is presented for measuring anisotropic THz response for small crystals, Crystal Anisotropy Terahertz Microscopy (CATM). Sucrose CATM measurements find the expected anisotropic phonon resonances. CATM measurements of protein crystals find the expected broadband water absorption is suppressed and strong orientation and hydration dependent resonant features.
Methods for patterning biomolecules on a substrate at the single molecule level have been studied as a route to sensors with single-molecular sensitivity or as a way to probe biological phenomena at the single-molecule level. However, the arrangement and orientation of single biomolecules on substrates has been less investigated. Here, we examined the arrangement and orientation of two rod-like coiled-coil proteins, cortexillin and tropomyosin, around patterned gold nanostructures. The high aspect ratio of the coiled coils made it possible to study their orientations and to pursue a strategy o…
The molecular origins of proteins' functions are a combinatorial search problem in the proteins' sequence space, which requires enormous resources to solve. However, evolution has already solved this optimization problem for us, leaving behind suboptimal solutions along the way. Comparing suboptimal proteins along the evolutionary pathway, or ancestors, with more optimal modern proteins can lead us to the exact molecular origins of a particular function. In this paper, we study the long-standing question of the selectivity of Imatinib, an anti-cancer kinase inhibitor drug. We study two related…
Quantum mechanical calculations are performed on the proteins that constitute the ubiquitin-Dsk2 complex whose atomic structure has been experimentally determined by NMR spectroscopy (PDB id 1WR1). The results indicate that the dipole moment vectors of the two proteins are aligned in a head-to-tail orientation while forming and angle of ~130°. Hence, attractive dipole-dipole interactions not only stabilize the protein-protein complex but they are likely to favor the correct orientation of the proteins during the formation of the complex.
The beautiful structures of single and multi-domain proteins are clearly ordered in some fashion but cannot be readily classified using group theory methods that are successfully used to describe periodic crystals. For this reason, protein structures are considered to be aperiodic, and may have evolved this way for functional purposes, especially in instances that require a combination of softness and rigidity within the same molecule. By analyzing the solved protein structures, we show that orientational symmetry is broken in the aperiodic arrangement of the secondary structural elements (SSE…
We investigate the fine structure of the sp3 hybridized covalent bond geometry that governs the tetrahedral architecture around the central C$_α$ carbon of a protein backbone, and for this we develop new visualization techniques to analyze high resolution X-ray structures in Protein Data Bank. We observe that there is a correlation between the deformations of the ideal tetrahedral symmetry and the local secondary structure of the protein. We propose a universal coarse grained energy function to describe the ensuing side-chain geometry in terms of the C$_β$ carbon orientations. The energy funct…
We present a new method to extract distance and orientation dependent potentials between amino acid side chains using a database of protein structures and the standard Boltzmann device. The importance of orientation dependent interactions is first established by computing orientational order parameters for proteins with alpha-helical and beta-sheet architecture. Extraction of the anisotropic interactions requires defining local reference frames for each amino acid that uniquely determine the coordinates of the neighboring residues. Using the local reference frames and histograms of the radial …
The past decade has witnessed the development and success of coarse-grained network models of proteins for predicting many equilibrium properties related to collective modes of motion. Curiously, the results are usually robust towards the different cutoff distances used for constructing the residue networks from the knowledge of the experimental coordinates. In this study, we present a systematical study of network construction, and their effect on the predicted properties. Probing bond orientational order around each residue, we propose a natural partitioning of the interactions into an essen…
Protein data bank entries obtain distinct, reproducible flexibility characteristics determined by normal mode analyses of their three dimensional coordinate files. We study the effectiveness and sensitivity of this technique by analyzing the results on one class of glycosidases: family 10 xylanases. A conserved tryptophan that appears to affect access to the active site can be in one of two conformations according to X-ray crystallographic electron density data. The two alternate orientations of this active site tryptophan lead to distinct flexibility spectra, with one orientation thwarting th…
In the template-assistance model, normal prion protein (PrPC), the pathogenic cause of prion diseases such as Creutzfeldt-Jakob (CJD) in human, Bovine Spongiform Encephalopathy (BSE) in cow, and scrapie in sheep, converts to infectious prion (PrPSc) through an autocatalytic process triggered by a transient interaction between PrPC and PrPSc. Conventional studies suggest the S1-H1-S2 region in PrPC to be the template of S1-S2 $β$-sheet in PrPSc, and the conformational conversion of PrPC into PrPSc may involve an unfolding of H1 in PrPC and its refolding into the $β$-sheet in PrPSc. Here we cond…
Learning meaningful protein representation is important for a variety of biological downstream tasks such as structure-based drug design. Having witnessed the success of protein sequence pretraining, pretraining for structural data which is more informative has become a promising research topic. However, there are three major challenges facing protein structure pretraining: insufficient sample diversity, physically unrealistic modeling, and the lack of protein-specific pretext tasks. To try to address these challenges, we present the 3D Geometric Pretraining. In this paper, we propose a unifie…
A 3D pocket-aware and evolutionary conserved interaction guided diffusion model for molecular optimization
School of Computer Science and Engineering, Sun Yat-sen University（中山大学计算机科学与工程学院）；School of Pharmaceutical Science, Sun Yat-sen University（中山大学药学院）；West China School of Pharmacy, Sichuan University（四川大学华西药学院）
Zhejiang University（浙江大学）；Westlake University（西湖大学）；Westlake Institute for Advanced Study（西湖高等研究院）
The linear sequence of amino acids determines protein structure and function. Protein design, known as the inverse of protein structure prediction, aims to obtain a novel protein sequence that will fold into the defined structure. Recent works on computational protein design have studied designing sequences for the desired backbone structure with local positional information and achieved competitive performance. However, similar local environments in different backbone structures may result in different amino acids, indicating that protein structure's global context matters. Thus, we propose t…
Ligand-Conditioned Discrete Diffusion for Protein Sequence-Structure Co-Design
Xi’an University of Posts & Telecommunications（西安邮电大学）；National University of Singapore（新加坡国立大学）；Xi’an Jiaotong University（西安交通大学）；Institute of Systems Medicine, Chinese Academy of Medical Sciences（中国医学科学院系统医学研究院）
The ability to computationally generate novel yet physically foldable protein structures could lead to new biological discoveries and new treatments targeting yet incurable diseases. Despite recent advances in protein structure prediction, directly generating diverse, novel protein structures from neural networks remains difficult. In this work, we present a new diffusion-based generative model that designs protein backbone structures via a procedure that mirrors the native folding process. We describe protein backbone structure as a series of consecutive angles capturing the relative orientat…
Massive molecular simulations of drug-target proteins have been used as a tool to understand disease mechanism and develop therapeutics. This work focuses on learning a generative neural network on a structural ensemble of a drug-target protein, e.g. SARS-CoV-2 Spike protein, obtained from computationally expensive molecular simulations. Model tasks involve characterizing the distinct structural fluctuations of the protein bound to various drug molecules, as well as efficient generation of protein conformations that can serve as an complement of a molecular simulation engine. Specifically, we …
Secondary structure elements of many protein families exhibit differential conservation on their opposing faces. Amphipathic helices and beta-sheets by definition possess this property, and play crucial functional roles. This type of evolutionary trajectory of a protein family is usually critical to the functions of the protein family, as well as in creating functions within subfamilies. That is, differential conservation maintains properties of a protein structure related to its orientation, and that are important in packing, recognition, and catalysis. Here I define and formulate a new conce…
Simple, Efficient and Scalable Structure-aware Adapter Boosts Protein Language Models
School of Information Science and Engineering, East China University of Science and Technology（华东理工大学信息科学与工程学院）；Institute of Natural Sciences, Shanghai Jiao Tong University（上海交通大学自然科学学院）
Zhejiang University（浙江大学）；Alibaba-Zhejiang University Joint Research Institute of Frontier Technologies（阿里巴巴-浙江大学前沿技术联合研究院）
Self-supervised protein language models have proved their effectiveness in learning the proteins representations. With the increasing computational power, current protein language models pre-trained with millions of diverse sequences can advance the parameter scale from million-level to billion-level and achieve remarkable improvement. However, those prevailing approaches rarely consider incorporating knowledge graphs (KGs), which can provide rich structured knowledge facts for better protein representations. We argue that informative biology knowledge in KGs can enhance protein representation…
