The generative capacity of probabilistic protein sequence models.
The generative capacity of probabilistic protein sequence models.
复制标题
DOI：
10.1038/s41467-021-26529-9
复制
发表时间：
2021-11-02
影响因子：
16.6
通讯作者：
Haldane A
中科院分区：
文献类型：
--
作者：
McGee F;Hauri S;Novinger Q;Vucetic S;Levy RM;Carnevale V;Haldane A
Potts模型和变分自编码器（VAEs）最近作为生成蛋白序列模型（GPSMs）获得了广泛的应用，用于探索适应度景观和预测突变效应。尽管取得了令人鼓舞的结果，但目前的模型评估指标尚不清楚gpms是否忠实地再现了由于上位性而在自然序列中观察到的复杂的多残基突变模式。在这里，我们开发了一套序列统计来评估当前三种gpsm的“生成能力”：成对Potts hamilton模型、VAE模型和站点无关模型。我们表明，Potts模型的生成能力最大，因为模型生成的高阶突变统计量与自然序列的观测值一致，而VAE介于Potts模型和地点无关模型之间。重要的是，我们的工作为评估和解释GPSM准确性提供了一个新的框架，该框架强调了高阶协变和上位性的作用，对一般的概率序列模型具有更广泛的意义。生成模型在蛋白质设计中越来越受欢迎，但缺乏严格的指标来比较这些模型。在这里，作者提出了一组这样的指标，并用它们来比较三种流行的模型。
Potts models and variational autoencoders (VAEs) have recently gained popularity as generative protein sequence models (GPSMs) to explore fitness landscapes and predict mutation effects. Despite encouraging results, current model evaluation metrics leave unclear whether GPSMs faithfully reproduce the complex multi-residue mutational patterns observed in natural sequences due to epistasis. Here, we develop a set of sequence statistics to assess the “generative capacity” of three current GPSMs: the pairwise Potts Hamiltonian, the VAE, and the site-independent model. We show that the Potts model’s generative capacity is largest, as the higher-order mutational statistics generated by the model agree with those observed for natural sequences, while the VAE’s lies between the Potts and site-independent models. Importantly, our work provides a new framework for evaluating and interpreting GPSM accuracy which emphasizes the role of higher-order covariation and epistasis, with broader implications for probabilistic sequence models in general. Generative models have become increasingly popular in protein design, yet rigorous metrics that allow the comparison of these models are lacking. Here, the authors propose a set of such metrics and use them to compare three popular models.
登录
查看更多内容
影响因子：
32.4
作者：
Ferguson AL;Mann JK;Omarjee S;Ndung'u T;Walker BD;Chakraborty AK
通讯作者：
Chakraborty AK
影响因子：
7.7
作者：
Biswas, Avik;Haldane, Allan;Levy, Ronald M.
通讯作者：
Levy, Ronald M.
影响因子：
4.3
作者：
Facco, Elena;Pagnani, Andrea;Laio, Alessandro
通讯作者：
Laio, Alessandro
DOI：
10.1073/pnas.1702664114
发表时间：
2017-08-22
影响因子：
11.1
作者：
Anishchenko, Ivan;Ovchinnikov, Sergey;Baker, David
通讯作者：
Baker, David
影响因子：
14.9
作者：
UniProt Consortium
通讯作者：
UniProt Consortium