超越本源序列恢复:改善了蛋白质结构的序列能量格局的建模
bioRxiv : the preprint server for biology
|February 6, 2026
概括
在蛋白质设计中优化机器学习模型以实现本机序列恢复 (NSR) 是不够的. 像PottsMPNN这样的新方法通过专注于折叠兼容性来改善序列生成和能量预测,而不仅仅是NSR.
科学领域:
- 计算生物学是一种计算生物学.
- 蛋白质工程是一种蛋白质工程.
- 机器学习是机器学习.
背景情况:
- 机器学习模型,包括AlphaFold2,已经显著提升了计算蛋白质设计.
- 当前的蛋白质设计工作流程通常涉及生成序列以稳定设计的骨干.
- 原生序列恢复 (NSR) 是序列设计模型的常见培训目标.
研究的目的:
- 为了证明优化序列设计模型仅用于本机序列恢复 (NSR) 的局限性.
- 介绍和评估PottsMPNN,这是一个训练生成Potts能量功能的模型,用于改进蛋白质设计.
- 突出替代策略,增强序列生成和能源预测超越NSR.
主要方法:
- 训练有素的机器学习模型,重点是生成波茨能量函数 (PottsMPNN).
- 评估模型性能基于序列与所需折叠的兼容性和突变效应的预测,而不仅仅是NSR.
- 经过训练的PottsMPNN使用噪声的骨干结构和多个序列对齐来测试强度.
主要成果:
- 仅针对NSR进行优化被证明与关键性能指标不一致.
- 在Potts能源模型上训练的PottsMPNN减少了NSR,但改善了序列生成和能源预测.
- 使用噪声数据的训练降低了NSR,但提高了设计序列和能量预测的质量.
结论:
- 原生序列恢复 (NSR) 不是评估基于机器学习的蛋白质序列设计模型的最佳指标.
- 训练模型以产生波茨能量功能为设计稳定和功能蛋白序列提供了更有效的方法.
- 这项工作为蛋白质设计提供了新的方向,超越了NSR优化.
相关概念视频
Cis-regulatory Sequences
11.9K
Cis-regulatory sequences are short fragments of non-coding DNA that are present on the same chromosomes as the genes that they regulate. These fragments serve as binding sites for transcriptional regulators, proteins that are responsible for controlling gene transcription and differential gene expression across cell types in eukaryotes. Cis-regulatory sequences can be close to the gene of interest or thousands of bases away in the DNA sequence; however, those sequences that are further away are...
11.9K
Cis-regulatory Sequences
4.2K
4.2K
Per-Unit Sequence Models
453
An ideal Y-Y transformer, grounded through neutral impedances, displays per-unit sequence networks akin to those of a single-phase ideal transformer when subjected to balanced positive- or negative-sequence currents. These currents do not produce neutral currents, and their associated voltage drops.
Zero-sequence currents, which are identical in magnitude and phase, generate a neutral current, resulting in voltage drops across the neutral impedance and the low-voltage winding. If the...
Zero-sequence currents, which are identical in magnitude and phase, generate a neutral current, resulting in voltage drops across the neutral impedance and the low-voltage winding. If the...
453
Sequences
280
Sequences are fundamental mathematical objects consisting of ordered lists of numbers that follow a specific rule or pattern. Sequences are critical in various mathematical concepts, including calculus, series, and number theory. They can model real-world phenomena such as population growth, financial investments, and physical processes like the diminishing height of a bouncing ball.Each number in a sequence is referred to as a term. Typically, the terms are denoted as a1, a2, a3,…, where...
280
Protein and Protein Structure
88.2K
Proteins are one of the most abundant organic molecules in living systems and have the most diverse range of functions of all macromolecules. Proteins may be structural, regulatory, contractile, or protective. They may serve in transport, storage, or membranes; or they may be toxins or enzymes. Their structures, like their functions, vary greatly. They are all, however, amino acid polymers arranged in a linear sequence.
A protein's shape is critical to its function. For example, an enzyme...
A protein's shape is critical to its function. For example, an enzyme...
88.2K
Sanger Sequencing
774.7K
DNA sequencing is a fundamental technique that is routinely used in the biological sciences. This method can be applied to a range of questions at different scales - from the sequencing of a cloned DNA fragment or the study of a mutation in a gene up to whole-genome sequencing. However, despite the widespread use of sequencing today, it was not until 1977 that Fredrick Sanger and his collaborators developed the chain-termination method to decode DNA sequences. It relies on the separation of a...
774.7K


