|
|
本帖最后由 邓文龙 于 2021-12-12 11:20 编辑
深度学习算法设计全新蛋白质
通过深度网络预测从头设计蛋白质
环球科学
2021/12/09
论文
论文标题：De novo protein design by deep network hallucination
作者：Anishchenko, Ivan, Pellock, Samuel J., Chidyausiku, Tamuka M., Ramelot, Theresa A., Ovchinnikov, Sergey, Hao, Jingzhou, Bafna, Khushboo, Norn, Christoffer, Kang, Alex, Bera, Asim K., DiMaio, Frank, Carter, Lauren, Chow, Cameron M., Montelione, Gaetano T., Baker, David
期刊：Nature
发表时间：2021/12/01
数字识别码：10.1038/s41586-021-04184-w
摘要：There has been considerable recent progress in protein structure prediction using deep neural networks to predict inter-residue distances from amino acid sequences1,2,3. Here we investigate whether the information captured by such networks is sufficiently rich to generate new folded proteins with sequences unrelated to those of the naturally occurring proteins used in training the models. We generate random amino acid sequences, and input them into the trRosetta structure prediction network to predict starting residue–residue distance maps, which, as expected, are quite featureless. We then carry out Monte Carlo sampling in amino acid sequence space, optimizing the contrast (Kullback–Leibler divergence) between the inter-residue distance distributions predicted by the network and background distributions averaged over all proteins. Optimization from different random starting points resulted in novel proteins spanning a wide range of sequences and predicted structures. We obtained synthetic genes encoding 129 of the network-‘hallucinated’ sequences, and expressed and purified the proteins in Escherichia coli; 27 of the proteins yielded monodisperse species with circular dichroism spectra consistent with the hallucinated structures. We determined the three-dimensional structures of three of the hallucinated proteins, two by X-ray crystallography and one by NMR, and these closely matched the hallucinated models. Thus, deep networks trained to predict native protein structures from their sequences can be inverted to design new proteins, and such networks and methods should contribute alongside traditional physics-based models to the de novo design of proteins with new functions.
所属学科：
计算机
生物
由多肽链折叠成的、具有复杂三维结构的蛋白质，是细胞发育、DNA修复、新陈代谢等几乎所有生物学活动的关键基础。设计、开发具有崭新序列和功能的蛋白质具有重要意义。然而，理解蛋白质的折叠和结构需要复杂的实验测算。近日，科学家基于深度学习算法开发了一个计算神经网络，它能帮助理解蛋白质的结构并预测任意多肽链的折叠方式，显著提高了蛋白质设计的效率，相关结果发表在《自然》上。
根据已知蛋白质的序列和结构，这个计算神经网络能预测研究人员设计的完全随机的氨基酸序列是否能形成稳定的空间结构。利用该计算神经网络，研究人员得到了2000多种可折叠成稳定结构的全新蛋白质，其中129种已经在实验室中合成并纯化出来，27种的二级结构与计算神经网络的预测一致。结合核磁共振、X射线晶体衍射等技术，科学家对合成的3种全新的蛋白质进行结构分析，发现它们的三维结构与预测非常相符。该研究展示了深度学习算法进行全新蛋白质结构预测的高准确性，开辟了蛋白质设计和改造的新途径。
文章标签
蛋白质结构
多肽链
折叠
序列
氨基酸序列
计算神经网络
深度学习算法
https://www.nature.com/articles/s41586-021-04184-w
https://www.linkresearcher.com/t ... c-8108-906431c8ff18
华成旅行社 欢迎来电咨询：
电话：03-3833-9823 / 03-5688-1863
FAX ：03-3833-9873 / 03-3834-5891
SOFTBANK电话：080-3416-2275 担当：小郭 微信号：08034162275
SOFTBANK电话：090-2172-4325 担当：小于 微信号：TYOSCL4325
SOFTBANK电话：080-3398-4387 担当：小李 微信号：huacheng4387
SOFTBANK电话：080-3523-4388 担当：小何 微信号：huacheng602
SOFTBANK电话：080-3084-4389 担当：小马 微信号：huacheng858
http://www.kaseisyoji.com/forum.php?mod=forumdisplay&fid=10
|
|
IP卡
狗仔卡
发表于 2021-12-12 11:10:06
收藏
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡
显身卡
楼主
