相关实验视频

Updated: Jan 16, 2026

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

国际PLM:通过稀疏的自编码器发现蛋白质语言模型中的可解释特征

Elana Simon¹, James Zou²

¹Stanford University, Stanford, CA, USA. epsimon@stanford.edu.

|September 30, 2025

概括

我们开发了一个框架,从蛋白质语言模型 (PLM) 中提取可解释的生物特征. 这种方法揭示了像ESM-2这样的PLM如何编码概念,帮助蛋白质设计和注释.

科学领域:

计算生物学是一种计算生物学.
在生物信息学中的机器学习.
蛋白质信息学是指蛋白质信息学.

背景情况:

蛋白质语言模型 (PLM) 在蛋白质建模和设计方面表现出色,但它们的内部运作仍然不清楚.
了解PLM机制对于推进蛋白质科学和工程至关重要.
解释像PLM这样的复杂模型是生物信息学中的一个重大挑战.

研究的目的:

开发一个系统的框架,从PLM中提取和分析可解释的特征.
调查PLM嵌入式中生物概念是如何表示的.
在蛋白质注释和生成中展示可解释特征的实际应用.

主要方法:

利用稀疏的自编码器来分析ESM-2蛋白语言模型中的嵌入.
训练稀疏的自动编码器来识别和提取可解释的特征.
采用大型语言模型进行自动化功能描述和验证.
在不同的模型尺度上研究了特征表示.

主要成果:

识别了数千个与生物概念相对应的可解释特征 (例如,结合点,图案,领域).
发现概念存储在PLM内的神经元之间叠加.
观察到,较大的PLM捕捉了更多的可解释概念.
发现ESM-2在已知注释之外的多种蛋白质家族中学习模式.
在识别缺失的数据库注释和指导序列生成方面证明了功能实用性.

结论:

PLM表示可以分解成有意义的,可解释的组件.
开发的框架提供了一种可行的和有用的方法来机械地解释PLM.
这项工作增强了我们对PLM的理解,并为蛋白质设计和发现开辟了新的途径.

更多相关视频

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

Author Spotlight: A Computational Approach to Decipher Amino Acid Preferences in Multispecific Protein-Protein Interactions

Author Spotlight: A Computational Approach to Decipher Amino Acid Preferences in Multispecific Protein-Protein Interactions

Published on: January 26, 2024

相关概念视频

Improving Translational Accuracy

Improving Translational Accuracy

Improving Translational Accuracy

Improving Translational Accuracy

Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...

Conservation of Protein Domains Over Different Proteins

Conservation of Protein Domains Over Different Proteins

Protein domains are small structurally independent units that are part of a single amino acid chain. Although these domains are often structurally independent, they may rely on synergistic effects to perform their functions as part of a larger protein. Protein domains may be conserved within the same organism, as well as across different organisms.
A limited set of protein domains often duplicate and recombine during evolution. These domains can be organized in different combinations to...

Conservation of Protein Domains

Conservation of Protein Domains

Protein Networks

Protein Networks

Protein-protein Interfaces

Protein-protein Interfaces

Many proteins form complexes to carry out their functions, making protein-protein interactions (PPIs) essential for an organism's survival. Most PPIs are stabilized by numerous weak noncovalent chemical forces. The physical shape of the interfaces determines the way two proteins interact. Many globular proteins have closely-matching shapes on their surfaces, which form a large number of weak bonds. Additionally, many PPIs occur between two helices or between a surface cleft and a...