Auto Byte

专注未来出行及智能汽车科技

微信扫一扫获取更多资讯

Science AI

关注人工智能与其他前沿技术、基础学科的交叉研究与融合发展

微信扫一扫获取更多资讯

ScienceAI 2021 蛋白质专题年度回顾

编辑 | 萝卜皮

毫无疑问,2021 年人类在蛋白质领域的探索取得了前所未有的成就。

在 2021 年 7 月 15 日,谷歌的 DeepMind 团队以及华盛顿大学的 Baker 团队,分别在 Nature 和 Science 两本顶刊杂志上发布并开源了蛋白质结构预测工具 AlphaFold2 与 RoseTTAFold。

这对于蛋白质领域是无疑是重磅进展。在之后的研究中,来自世界各地的科学家团队使用开源的工具对未知蛋白进行探索,同时也不断验证了新工具的稳健性。更有团队以开源的蛋白质结构预测工具为基准,开发了蛋白互作预测工具。

ScienceAI 认为,2021年蛋白质结构预测所取得的进展,是分子生物学发展的一座里程碑。

以下是 ScienceAI 2021 年蛋白质专题报道的年度总结。

友情提示:点击小标题可直接访问文章

一天之内,两大 AI 预测蛋白结构算法开源,分别登上 Nature、Science

Nature | Highly accurate protein structure prediction with AlphaFold

论文链接:https://www.nature.com/articles/s41586-021-03819-2

开源地址:https://github.com/deepmind/alphafold

Science | Accurate prediction of protein structures and interactions using a three-track neural network

论文链接:https://science.sciencemag.org/content/early/2021/07/14/science.abj8754

开源地址:https://github.com/RosettaCommons/RoseTTAFold


AlphaFold 再登 Nature!预测确定 98.5% 所有人类蛋白结构,数据库全部免费开放

Nature | Highly accurate protein structure prediction for the human proteome

论文链接:https://www.nature.com/articles/s41586-021-03828-1


人工智能为蛋白质折叠预测提供动力

Nature | Artificial intelligence powers protein-folding predictions

论文链接:https://www.nature.com/articles/d41586-021-03499-y


生物计算专家超细致解读 AlphaFold2 论文:模型架构及应用

来自哥伦比亚大学欧文医学中心 Mohammed AlQuraishi 的个人博客

原文链接:https://moalquraishi.wordpress.com/2021/07/25/the-alphafold2-method-paper-a-fount-of-good-ideas/


网红教授发 Nature 品评 AlphaFold2,蛋白质结构预测将彻底改变

Nature | Protein-structure prediction revolutionized

论文链接:https://www.nature.com/articles/d41586-021-02265-4


AlphaFold 2 对蛋白结构研究领域的冲击有多大,听听这五位专家怎么说

来自机器之心策划的《AlphaFold 2「能」与「不能」》知识分享活动

视频回放链接:https://jmq.h5.xeknow.com/s/2ZtoeT


Nature社论:AlphaFold 为生命科学带来了什么?开源的技术和未来的方向

Nature | Artificial intelligence in structural biology is here to stay

论文链接:https://www.nature.com/articles/d41586-021-02037-0


借助 AlphaFold2 对噬菌体粘附装置的结构和拓扑进行解析

Microorganisms | Structure and Topology Prediction of Phage Adhesion Devices Using AlphaFold2: The Case of Two Oenococcus oeni Phages

论文链接:https://www.mdpi.com/2076-2607/9/10/2151/htm


已开源新工具:在 AlphaFold 建模的基础上,解析蛋白质的糖基化修饰

Nature Structural & Molecular Biology | The case for post-predictional modifications in the AlphaFold Protein Structure Database

数据链接:https://zenodo.org/record/5564681

论文链接:https://www.nature.com/articles/s41594-021-00680-9


人工智能揭示核孔结构,再次证明 AlphaFold 和 RoseTTAfold 预测的可靠性

bioRxiv 预印平台 | Artificial intelligence reveals nuclear pore complexity

论文链接:https://www.biorxiv.org/content/10.1101/2021.10.26.465776v1.full.pdf


预测结果与实验数据基本一致,AlphaFold2 应用于研究蛋白活化以及相互作用

Molecular Reproduction and Development | Using machine learning to study protein–protein interactions: From the uromodulin polymer to egg zona pellucida filaments

论文链接:https://onlinelibrary.wiley.com/doi/10.1002/mrd.23538


蛋白结构预测与结构生物学的未来

Nature Structural & Molecular Biology | AlphaFold2 and the future of structural biology

论文链接:https://www.nature.com/articles/s41594-021-00650-1

Nature Biotechnology | The 3D protein deluge

论文链接:https://www.nature.com/articles/s41587-021-01029-9


无需「协同进化」信息,芝加哥许锦波团队最新研究登上Nature子刊

Nature Machine Intelligence | Improved protein structure prediction by deep learning irrespective of co-evolution information

论文链接:https://www.nature.com/articles/s42256-021-00348-5


只需1台CPU跑600多秒,GNN快速有效优化蛋白质模型

Nature Computational Science | Fast and effective protein model refinement using deep graph neural networks

开源地址:http://raptorx.uchicago.edu/

论文链接:https://www.nature.com/articles/s43588-021-00098-9


深度学习解析蛋白-蛋白互作关系,助力细胞代谢途径研究

Science | Computed structures of core eukaryotic protein complexes

论文链接:https://www.science.org/doi/10.1126/science.abm4805


已开源,多个国内团队使用人工智能揭示蛋白质相互作用

Bioinformatics | PhosIDN: an integrated deep neural network for improving protein phosphorylation site prediction by combining sequence and protein–protein interaction information

开源地址:https://github.com/ustchangyuanyang/PhosIDN

论文链接:https://academic.oup.com/bioinformatics/article/37/24/4668/6329824

Bioinformatics | Transfer learning via multi-scale convolutional neural layers for human–virus protein–protein interaction prediction

开源地址:https://github.com/XiaodiYangCAU/TransPPI/

论文链接:https://academic.oup.com/bioinformatics/article-abstract/37/24/4771/6323357?redirectedFrom=fulltext

Bioinformatics | DeepTrio: a ternary prediction system for protein–protein interaction using mask multiple parallel convolutional neural networks

开源地址:https://github.com/huxiaoti/deeptrio.git

论文链接:https://academic.oup.com/bioinformatics/advance-article-abstract/doi/10.1093/bioinformatics/btab737/6409848?redirectedFrom=fulltext

Bioinformatics | pyconsFold: a fast and easy tool for modeling and docking using distance predictions

开源地址:https://github.com/johnlamb/pyconsfold

论文链接:https://academic.oup.com/bioinformatics/article/37/21/3959/6317824

Nature Machine Intelligence | A geometric deep learning approach to predict binding conformations of bioactive molecules

论文链接:https://www.nature.com/articles/s42256-021-00409-9


8张RTX3090,效果媲美AlphaFold2,国产蛋白结构预测平台TRFold排名全球第二

来自对天壤团队的采访


绘制人体蛋白质图谱的人类蛋白质组计划

Science Advances | The Human Proteoform Project: Defining the human proteome

论文链接:https://www.science.org/doi/10.1126/sciadv.abk0734


集成「进化环境」的深度学习算法加速蛋白质工程

Nature Communications | ECNet is an evolutionary context-integrated deep learning framework for protein engineering

论文链接:https://www.nature.com/articles/s41467-021-25976-8


神经网络学习预测蛋白「分子机器」的运动

Nature Methods | Neural networks learn the motions of molecular machines

论文链接:https://www.nature.com/articles/s41592-021-01235-y


科学家利用深度神经网络的想象,进行从头蛋白质设计

Nature | De novo protein design by deep network hallucination

论文链接:https://www.nature.com/articles/s41586-021-04184-w


利用进化深度卷积神经网络识别全基因组RNA结合蛋白

Bioinformatics | EDCNN: identification of genome-wide RNA-binding proteins using evolutionary deep convolutional neural network

论文链接:https://academic.oup.com/bioinformatics/advance-article-abstract/doi/10.1093/bioinformatics/btab739/6409850?redirectedFrom=fulltext


人工智能发现人类蛋白质具有杀灭超级细菌的潜力

Nature Biomedical Engineering | Mining for encrypted peptide antibiotics in the human proteome

论文链接:https://www.nature.com/articles/s41551-021-00801-1


上科大研究登Nature子刊,深度学习更快、更深入地进行磷酸化蛋白质组分析

Nature Communications | DeepPhospho accelerates DIA phosphoproteome profiling through in silico library generation

论文链接:https://www.nature.com/articles/s41467-021-26979-1


AI 从底物和酶的结构中预测米氏常数,量化酶活性

PLOS BIOLOGY | Deep learning allows genome-scale prediction of Michaelis constants from structural features

论文链接:https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001402


纳米孔检测特定蛋白质,使细胞能够与计算机通话

Nature Biotechnology | Multiplexed direct detection of barcoded protein reporters on a nanopore array

论文链接:https://www.nature.com/articles/s41587-021-01002-6


人工智能对「可疑」蛋白质进行「指纹识别」

PNAS | Single-particle diffusional fingerprinting: A machine-learning framework for quantitative analysis of heterogeneous diffusion

论文链接:https://www.pnas.org/content/118/31/e2104624118#abstract-2


使用 3D 卷积神经网络检测蛋白质-肽结合位点,以启动肽药物发现

Journal of Chemical Information and Modeling | Protein−Peptide Binding Site Detection Using 3D Convolutional Neural Networks

论文链接:https://pubs.acs.org/doi/10.1021/acs.jcim.1c00475


AI伦理领袖、AlphaFold作者上榜!Nature发布「2021十大科学人物」

原文链接:https://www.nature.com/immersive/d41586-021-03621-0/index.html#section-wEfZ0Jxqu4


Science封面 | 蛋白质结构预测大放异彩

原文链接:https://www.science.org/content/article/breakthrough-2021#section_breakthrough

理论蛋白质
暂无评论
暂无评论~