检索结果-内蒙古大学图书馆

Modeling global dynamics from local snapshots with deep generative neural networks

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Gigante, Scott Van Dijk, David Moon, Kevin R. Strzalkowski, Alexander Wolf, Guy Krishnaswamy, Smita Computational Biology and Bioinformatics Program Yale University New HavenCT United States Department of Genetics Yale University New HavenCT United States Department of Computer Science Yale University New HavenCT United States Department of Mathematics and Statistics Utah State University LoganUT United States Department of Computer Science Princeton University PrincetonNJ United States Department of Mathematics and Statistics Universitéde Montreacuteal MontéalQC Canada

Complex high dimensional stochastic dynamic systems arise in many applications in the natural sciences and especially biology. However, while these systems are difficult to describe analytically, "snapshot" measurements that sample the output of the system are often available. In order to model the dynamics of such systems given snapshot data, or local transitions, we present a deep neural network framework we call Dynamics Modeling Network or DyMoN. DyMoN is a neural network framework trained as a deep generative Markov model whose next state is a probability distribution based on the current state. DyMoN is trained using samples of current and nextstate pairs, and thus does not require longitudinal measurements. We show the advantage of DyMoN over shallow models such as Kalman filters and hidden Markov models, and other deep models such as recurrent neural networks in its ability to embody the dynamics (which can be studied via perturbation of the neural network) and generate longitudinal hypothetical trajectories. We perform three case studies in which we apply DyMoN to different types of biological systems and extract features of the dynamics in each case by examining the learned model. Copyright © 2018, The Authors. All rights reserved.

关键词： Deep neural networks

A Survey of Generative AI for de novo Drug Design: New Frontiers in Molecule and Protein Generation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Tang, Xiangru Dai, Howard Knight, Elizabeth Wu, Fang Li, Yunyang Li, Tianxiao Gerstein, Mark Department of Computer Science Yale University New HavenCT06520 United States School of Medicine Yale University New HavenCT06520 United States Computer Science Department Stanford University CA94305 United States Program in Computational Biology & Bioinformatics Yale University New HavenCT06520 United States Department of Molecular Biophysics & Biochemistry Yale University New HavenCT06520 United States Department of Statistics & Data Science Yale University New HavenCT06520 United States Department of Biomedical Informatics & Data Science Yale University New HavenCT06520 United States

Artificial intelligence (AI)-driven methods can vastly improve the historically costly drug design process, with various generative models already in widespread use. Generative models for de novo drug design, in particular, focus on the creation of novel biological compounds entirely from scratch, representing a promising future direction. Rapid development in the field, combined with the inherent complexity of the drug design process, creates a difficult landscape for new researchers to enter. In this survey, we organize de novo drug design into two overarching themes: small molecule and protein generation. Within each theme, we identify a variety of subtasks and applications, highlighting important datasets, benchmarks, and model architectures and comparing the performance of top models. We take a broad approach to AI-driven drug design, allowing for both micro-level comparisons of various methods within each subtask and macro-level observations across different fields. We discuss parallel challenges and approaches between the two applications and highlight future directions for AI-driven de novo drug design as a whole. An organized repository of all covered sources is available at https://***/gersteinlab/GenAI4Drug. © 2024, CC BY.

关键词： Proteins

Improved Prediction Of Ligand-Protein Binding Affinities By Meta-Modeling

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Lee, Ho-Joon Emani, Prashant S. Gerstein, Mark B. Dept. of Genetics and Yale University New HavenCT06510 United States Yale Center for Genome Analysis Yale University New HavenCT06510 United States Dept. of Molecular Biophysics & Biochemistry Yale University New HavenCT06520 United States Program in Computational Biology & Bioinformatics Yale University New HavenCT06520 United States Dept. of Computer Science Yale University New HavenCT06520 United States Dept. of Statistics & Data Science Yale University New HavenCT06520 United States Dept. of Biomedical Informatics & Data Science Yale University New HavenCT06520 United States

The accurate screening of candidate drug ligands against target proteins through computational approaches is of prime interest to drug development efforts. Such virtual screening depends in part on methods to predict the binding affinity between ligands and proteins. Many computational models for binding affinity prediction have been developed, but with varying results across targets. Given that ensembling or meta-modeling approaches have shown great promise in reducing model-specific biases, we develop a framework to integrate published force-field-based empirical docking and sequence-based deep learning models. In building this framework, we evaluate many combinations of individual base models, training databases, and several meta-modeling approaches. We show that many of our meta-models significantly improve affinity predictions over base models. Our best meta-models achieve comparable performance to state-of-the-art deep learning tools exclusively based on 3D structures, while allowing for improved database scalability and flexibility through the explicit inclusion of features such as physicochemical properties or molecular descriptors. We further demonstrate improved generalization capability by our models using a large-scale benchmark of affinity prediction as well as a virtual screening application benchmark. Overall, we demonstrate that diverse modeling approaches can be ensembled together to gain meaningful improvement in binding affinity prediction. © 2023, CC BY-NC-SA.

关键词： Prediction models

Distinct modes of mitochondrial metabolism uncouple T cell differentiation and function (vol 571, pg 403, 2019)

学校读者我要写书评

暂无评论

NATURE 2019年第7773期573卷 E2-E2页

作者： Bailis, Will Shyer, Justin A. Zhao, Jun Canaveras, Juan Carlos Garcia Al Khazal, Fatimah J. Qu, Rihao Steach, Holly R. Bielecki, Piotr Khan, Omair Jackson, Ruaidhri Kluger, Yuval Maher, Louis J., III Rabinowitz, Joshua Craft, Joe Flavell, Richard A. Department of Immunobiology Yale School of Medicine New Haven USA Department of Pathology Children’s Hospital of Philadelphia Philadelphia USA Department of Pathology Yale School of Medicine New Haven USA Program of Computational Biology and Bioinformatics Yale University New Haven USA Lewis-Sigler Institute for Integrative Genomics Princeton University Princeton USA Department of Chemistry Princeton University Princeton USA Diabetes Research Center University of Pennsylvania Philadelphia USA Department of Biochemistry and Molecular Biology Mayo Clinic College of Medicine and Science Rochester USA Applied Mathematics Program Yale University New Haven USA Department of Internal Medicine (Rheumatology) Yale School of Medicine New Haven USA Howard Hughes Medical Institute Chevy Chase USA

An Amendment to this paper has been published and can be accessed via a link at the top of the paper.

关键词：

Universal mechanical response of metallic glasses during strain-rate-dependent uniaxial compression

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Jin, Weiwei Datye, Amit Schwarz, Udo D. Shattuck, Mark D. O'Hern, Corey S. Department of Mechanical Engineering and Materials Science Yale University New HavenCT06520 United States Department of Chemical and Environmental Engineering Yale University New HavenCT06520 United States Benjamin Levich Institute Physics Department The City College of New York New YorkNY10031 United States Department of Physics Yale University New HavenCT06520 United States Department of Applied Physics Yale University New HavenCT06520 United States Graduate Program in Computational Biology and Bioinformatics Yale University New HavenCT06520 United States

Experimental data on the compressive strength σmax versus strain rate Ε eng for metallic glasses undergoing uniaxial compression shows significantly different behavior for different alloys. For some metallic glasses, σmax decreases with increasing Ε eng, for others, σmax increases with increasing Ε eng, and for others σmax versus Ε eng is nonmonotonic. Using numerical simulations of metallic glasses undergoing uniaxial compression at nonzero strain rate and temperature, we show that they obey a universal relation for the compressive strength versus temperature, which determines their mechanical response. At low Ε eng, increasing strain rate leads to increases in temperature and decreases in σmax∗, whereas at high Ε eng, increasing strain rate leads to decreases in temperature and increases in σmax∗. This non-monotonic behavior of σmax∗ versus temperature causes the nonmonotonic behavior of σmax∗ versus Ε eng. Variations in the internal dissipation change the characteristic strain rate at which the nonmonotonic behavior occurs. These results are general for a wide range of metallic glasses with different atomic interactions, damping coefficients, and chemical compositions. © 2022, CC BY.

关键词： Strain rate

Hypostatic jammed packings of frictionless nonspherical particles

学校读者我要写书评

暂无评论

arXiv 2017年

作者： VanderWerf, Kyle Jin, Weiwei Shattuck, Mark D. O'Hern, Corey S. Department of Physics Yale University New HavenCT06520 United States Benjamin Levich Institute and Physics Department City College of New York New YorkNY10031 United States Department of Mechanics and Engineering Science Peking University Beijing100871 China Department of Mechanical Engineering & Materials Science Yale University New HavenCT06520 United States Department of Applied Physics Yale University New HavenCT06520 United States Graduate Program in Computational Biology and Bioinformatics Yale University New HavenCT06520 United States

We perform computational studies of static packings of a variety of nonspherical particles including circulo-lines, circulo-polygons, ellipses, asymmetric dimers, dumbbells, and others to determine which shapes form packings with fewer contacts than degrees of freedom (hypostatic packings) and which have equal numbers of contacts and degrees of freedom (isostatic packings), and to understand why hypostatic packings of nonspherical particles can be mechanically stable despite having fewer contacts than that predicted from naïve constraint counting. To generate highly accurate force- and torque-balanced packings of circulo-lines and -polygons, we developed an interparticle potential that gives continuous forces and torques as a function of the particle coordinates. We show that the packing fraction and coordination number at jamming onset obey a master-like form for all of the nonspherical particle packings we studied when plotted versus the particle asphericity A, which is proportional to the ratio of the squared perimeter to the area of the particle. Further, the eigenvalue spectra of the dynamical matrix for packings of different particle shapes collapse when plotted at the same A. For hypostatic packings of nonspherical particles, we verify that the number of "quartic" modes along which the potential energy increases as the fourth power of the perturbation amplitude matches the number of missing contacts relative to the isostatic value. We show that the fourth derivatives of the total potential energy in the directions of the quartic modes remain nonzero as the pressure of the packings is decreased to zero. In addition, we calculate the principal curvatures of the inequality constraints for each contact in circulo-line packings and identify specific types of contacts with inequality constraints that possess convex curvature. These contacts can constrain multiple degrees of freedom and allow hypostatic packings of nonspherical particles to be mechanically stable. C

关键词： Degrees of freedom (mechanics)

Visualizing structure and transitions in high-dimensional biological data (vol 54, pg 781, 2019)

学校读者我要写书评

暂无评论

NATURE BIOTECHNOLOGY 2020年第1期38卷 108-108页

作者： Moon, Kevin R. van Dijk, David Wang, Zheng Gigante, Scott Burkhardt, Daniel B. Chen, William S. Yim, Kristina van den Elzen, Antonia Hirn, Matthew J. Coifman, Ronald R. Ivanova, Natalia B. Wolf, Guy Krishnaswamy, Smita Department of Mathematics and Statistics Utah State University Logan USA Cardiovascular Research Center section Cardiology Department of Internal Medicine Yale University New Haven USA Department of Computer Science Yale University New Haven USA School of Basic Medicine Qingdao University Qingdao China Yale Stem Cell Center Department of Genetics Yale University New Haven USA Computational Biology and Bioinformatics Program Yale University New Haven USA Department of Genetics Yale University New Haven USA Department of Computational Mathematics Science and Engineering Michigan State University East Lansing USA Department of Mathematics Michigan State University East Lansing USA Applied Mathematics Program Yale University New Haven USA Department of Genetics Center for Molecular Medicine University of Georgia Athens USA Department of Mathematics and Statistics Université de Montréal Montréal Canada Mila—Quebec Artificial Intelligence Institute Montréal Canada

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

关键词：

Graph Fourier MMD for Signals on Graphs

学校读者我要写书评

暂无评论

Graph Fourier MMD for Signals on Graphs

International Conference on Sampling Theory and Applications (SampTA)

作者： Samuel Leone Aarthi Venkat Guillaume Huguet Alexander Tong Guy Wolf Smita Krishnaswamy Applied Math Program New Haven Connecticut Computational Biology & Bioinformatics Program Yale University New Haven Connecticut Dept. of Math. & Stat Mila - Quebec AI Institute Université de Montréal Montreal Canada Dept. of Comp. Sci. & Oper. Res. Mila-Quebec AI Institute Université de Montréal Montreal Canada Department of Computer Science Department of Genetics Yale University New Haven Connecticut

While numerous methods have been proposed for computing distances between probability distributions in Euclidean space, relatively little attention has been given to computing such distances for distributions on graphs. However, there has been a marked increase in data that either lies on graph (such as protein interaction networks) or can be modeled as a graph (single cell data), particularly in the biomedical sciences. Thus, it becomes important to find ways to compare signals defined on such graphs. Here, we propose Graph Fourier MMD (GFMMD), a novel distance between distributions and signals on graphs. GFMMD is defined via an optimal witness function that is both smooth on the graph and maximizes the difference in expectation between the pair of distributions on the graph. We find an analytical solution to this optimization problem as well as an embedding of distributions that results from this method. We also prove several properties of this method including scale invariance and applicability to disconnected graphs. We showcase it on graph benchmark datasets as well on single cell RNA-sequencing data analysis. In the latter, we use the GFMMD-based gene embeddings to find meaningful gene clusters. We also propose a novel type of score for gene selection called gene localization score which helps select genes for cellular state space characterization.

关键词：

Current md forcefields fail to capture key features of protein structure and fluctuations: a case study of cyclophilin a and t4 lysozyme

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Mei, Zhe Grigas, Alex T. Treado, John D. Corres, Gabriel Melendez Vuorte, Maisa Sammalkorpi, Maria Regan, Lynne Levine, Zachary A. O’Hern, Corey S. Department of Chemistry Yale University New HavenCT06520 United States Integrated Graduate Program in Physical and Engineering Biology Yale University New HavenCT06520 United States Graduate Program in Computational Biology and Bioinformatics Yale University New HavenCT06520 United States Department of Mechanical Engineering and Materials Science Yale University New HavenCT06520 United States Department of Biology UPR Humacao Puerto RicoHumacao00792 Puerto Rico Department of Chemistry and Materials Science School of Chemical Engineering Aalto University Aalto Finland Department of Bioproducts and Biosystems School of Chemical Engineering Aalto University Aalto Finland Institute of Quantitative Biology Biochemistry and Biotechnology Center for Synthetic and Systems Biology School of Biological Sciences University of Edinburgh Edinburgh United Kingdom Department of Pathology Yale School of Medicine New HavenCT06520 United States Department of Molecular Biophysics and Biochemistry Yale University New HavenCT06520 United States Department of Physics Yale University New HavenCT06520 United States Department of Applied Physics Yale University New HavenCT06520 United States

Globular proteins undergo thermal fluctuations in solution, while maintaining an overall well-defined folded structure. In particular, studies have shown that the core structure of globular proteins differs in small, but significant ways when they are solved by x-ray crystallography versus solution-based NMR spectroscopy. Given these discrepancies, it is unclear whether molecular dynamics (MD) simulations can accurately recapitulate protein conformations. We therefore perform extensive MD simulations across multiple force fields and sampling techniques to investigate the degree to which computer simulations can capture the ensemble of conformations observed in experiments. By analyzing fluctuations in the atomic coordinates and core packing, we show that conformations sampled in MD simulations both move away from and sample a larger conformational space than the ensemble of structures observed in NMR experiments. However, we find that adding inter-residue distance restraints that match those obtained via Nuclear Overhauser Effect measurements enables the MD simulations to sample more NMR-like conformations, though significant differences between the core packing features in restrained MD and the NMR ensemble remain. Given that the protein structures obtained from the MD simulations possess smaller and less dense protein cores compared to those solved by NMR, we suggest that future improvements to MD forcefields should aim to increase the packing of hydrophobic residues in protein cores. © 2020, CC BY.

关键词： Molecular dynamics