We describe and test an implicit solvent all-atom potential for simulations of protein folding and aggregation. The potential is developed through studies of structural and thermodynamic properties of 17 peptides with diverse secondary structure. Results obtained using the final form of the potential are presented for all these peptides. The same model, with unchanged parameters, is furthermore applied to a heterodimeric coiled-coil system, a mixed alpha/beta protein and a three-helix-bundle protein, with very good results. The computational efficiency of the potential makes it possible to investigate the free-energy landscape of these 49--67-residue systems with high statistical accuracy, using only modest computational resources by today's standards.
We inquire to what extent protein peptide plane and side chain dynamics can be reconstructed from knowledge of C-alpha dynamics. Due to lack of experimental data we analyze all atom molecular dynamics trajectories from Anton supercomputer, and for clarity we limit our attention to the peptide plane O atoms and side chain C-beta atoms. We try and reconstruct their dynamics using four different approaches. Three of these are the publicly available reconstruction programs Pulchra, Remo Scwrl4. The fourth, Statistical Method, builds entirely on statistical analysis of Protein Data Bank (PDB) struc…
We study the behavior of five proteins at the air-water and oil-water interfaces by all-atom molecular dynamics. The proteins are found to get distorted when pinned to the interface. This behavior is consistent with the phenomenological way of introducing the interfaces in a coarse-grained model through a force that depends on the hydropathy indices of the residues. Proteins couple to the oil-water interface stronger than to the air- water one. They diffuse slower at the oil-water interface but do not depin from it, whereas depinning events are observed at the other interface. The reduction of…
Despite recent advancements in deep learning methods for protein structure prediction and representation, little focus has been directed at the simultaneous inclusion and prediction of protein backbone and sidechain structure information. We present SidechainNet, a new dataset that directly extends the ProteinNet dataset. SidechainNet includes angle and atomic coordinate information capable of describing all heavy atoms of each protein structure. In this paper, we provide background information on the availability of protein structure data and the significance of ProteinNet. Thereafter, we arg…
Here we propose that the upper bound marginal stability of proteins (7.4 kcal/mol) is a universal property that includes macro-molecular complexes and is not affected by molecular changes such as mutations and Post-Translational Modifications. Its existence is, essentially, a consequence of the Anfinsen thermodynamic hypothesis rather than a result of an evolutive process. This result enables us to conjecture that neutral evolution should also be, with respect to protein stability, a universal phenomenon.
ESM All-Atom: Multi-scale Protein Language Model for Unified Molecular Modeling
Department of Computer Science, Tsinghua University（清华大学计算机系）；Institute for AI Industry Research (AIR), Tsinghua University（清华大学人工智能产业研究院）；PharMolix Inc（PharMolix公司）
The rigidity and flexibility of homologous psychrophilic(P), mesophilic(M) and thermophilic(T) proteins have been investigated at the global and local levels in terms of packing factor and atomic fluctuations obtained from B-factors. For comparison of atomic fluctuations, correction of errors by considering errors in B-factors from all sources in a consolidated manner and conversion of the fluctuations to the same temperature have been suggested and validated. Results indicate no differences in the global values like average packing factor among the three classes of protein homologs but at loc…
One-Dimensional Structural Properties of Proteins in the Coarse-Grained CABS Model
尽管计算能力显著提高，但使用经典的全原子方法对蛋白质结构进行分子建模仍然效率低下，至少对于生物医学研究关注的大多数蛋白质靶点来说是如此。也许克服效率低下问题的最成功的策略是多尺度建模，以合并所有原子和粗粒度的模型。本章描述了一个成熟的CABS粗粒度蛋白质模型。CABS（C-Alpha，C-Beta and Side chains）模型假设氨基酸的2-4个联合原子表示、基于知识的力场（来自已知蛋白质序列和结构中的统计学数据）和有效的Monte Carlo采样方案（MC动力学、MC复制交换和组合）。特别强调的是CABS力场的独特设计，这在很大程度上是使用蛋白质的一维结构特性，包括蛋白质二级结构定义的。本章还介绍了基于CABS的建模方法，包括从头结构预测的多尺度工具，蛋白质动力学建模和蛋白质-肽复合物的预测。基于CABS的工具可在http://biocomp.chem.uw.edu.pl/tools上免费获得
Coarse-graining (CG) accelerates molecular simulations of protein dynamics by simulating sets of atoms as singular beads. Backmapping is the opposite operation of bringing lost atomistic details back from the CG representation. While machine learning (ML) has produced accurate and efficient CG simulations of proteins, fast and reliable backmapping remains a challenge. Rule-based methods produce poor all-atom geometries, needing computationally costly refinement through additional simulations. Recently proposed ML approaches outperform traditional baselines but are not transferable between prot…
Recently, persistent homology has had tremendous success in biomolecular data analysis. It works by examining the topological relationship or connectivity of a group of atoms in a molecule at a variety of scales, then rendering a family of topological representations of the molecule. However, persistent homology is rarely {employed} for the analysis of atomic properties, such as biomolecular flexibility analysis or B factor prediction. This work introduces atom-specific persistent homology to provide a local atomic level representation of a molecule via a global topological tool. This is achie…
Generative Modeling of Full-Atom Protein Conformations using Latent Diffusion on Graph Embeddings
Signal Processing Laboratory (LTS2), EPFL, Lausanne, Switzerland（信号处理实验室（LTS2），瑞士洛桑联邦理工学院）；Institute of Bioengineering, EPFL, Lausanne, Switzerland（生物工程研究所，瑞士洛桑联邦理工学院）；Bioinformatics Group, University & Research Wageningen, The Netherlands（生物信息学组，荷兰瓦赫宁根大学及研究学院）；Ludwig Institute for Cancer Research, Lausanne, Switzerland（癌症研究路易斯研究所，瑞士洛桑）
An Automatable Analytical Algorithm for Structure-Based Protein Functional Annotation via Detection of Specific Ligand 3D Binding Sites: Application to ATP (ser/thr Protein Kinases) and GTP (Small Ras-type G-Proteins) Binding Sites
Identifying novel functional protein structures is at the heart of molecular engineering and molecular biology, requiring an often computationally exhaustive search. We introduce the use of a Deep Convolutional Generative Adversarial Network (DCGAN) to classify protein structures based on their functionality by encoding each sample in a grid object structure using three features in each object: the generic atom type, the position atom type, and its occupancy relative to a given atom. We train DCGAN on 3-dimensional (3D) decoy and native protein structures in order to generate and discriminate …
Potential function of simplified protein models for discriminating native proteins from decoys: Combining contact interaction and local sequence-dependent geometry
Understanding and accurately predicting protein-ligand binding affinity are essential in the drug design and discovery process. At present, machine learning-based methodologies are gaining popularity as a means of predicting binding affinity due to their efficiency and accuracy, as well as the increasing availability of structural and binding affinity data for protein-ligand complexes. In biomolecular studies, graph theory has been widely applied since graphs can be used to model molecules or molecular complexes in a natural manner. In the present work, we upgrade the graph-based learners for …
Urea at sufficiently high concentration unfolds the secondary structure of proteins leading to denaturation. In contrast, Choline Chloride (ChCl) and urea, in 1:2 molar ratio form a deep eutectic mixture, a liquid at room temperature and protect proteins from denaturation. In order to get a microscopic picture of this phenomenon, we perform extensive all-atom molecular dynamics simulations on a model protein HP-36. Based on our calculation of Kirkwood-Buff integrals, we analyze the relative accumulation of these osmolytes around the protein. Additional insights are drawn from the analyses of t…
We report the results of our study of approved drugs as potential treatments for COVID 19, based on the application of various bioinformatics predictive methods. The drugs studied include chloroquine, ivermectin, remdesivir, sofosbuvir, boceprevir, and α-difluoromethylornithine (DMFO). Our results indicate that these small molecules selectively bind to stable, kinetically active residues and residues adjoining them on the surface of proteins and inside protein pockets and that some prefer hydrophobic over other active sites. Our approach is not restricted to viruses and can facilitate rational…
The Protein Data Bank (PDB) contains the atomic structures of over 105 biomolecules with better than 2.8A resolution. The listing of the identities and coordinates of the atoms comprising each macromolecule permits an analysis of the slow-time vibrational response of these large systems to minor perturbations. 3D video animations of individual modes of oscillation demonstrate how regions interdigitate to create cohesive collective motions, providing a comprehensive framework for and familiarity with the overall 3D architecture. Furthermore, the isolation and representation of the softest, slow…
