arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2405.15840q-bio.QMcs.LG
收藏

学习蛋白质结构的语言

Learning the Language of Protein Structure

InstaDeep(英斯塔迪普公司)

Benoit Gaujac, Jérémie Donà, Liviu Copoiu, Timothy Atkinson, Thomas Pierrot, Thomas D. Barrett

AI总结:

本文提出向量量化自编码器,将连续三维蛋白质结构离散为4096–64000个标记的码本,并以GPT生成新颖、可设计的结构,为蛋白质多模态建模奠定基础。

摘要:

蛋白质的表示学习与从头生成是计算生物学中的关键任务。尽管自然语言处理(NLP)技术已被证明在蛋白质序列建模方面非常有效,但结构建模却构成了一项复杂挑战,主要原因在于其连续且三维的特性。受这种差异的启发,我们提出一种使用向量量化自编码器的方法,可有效地将蛋白质结构标记化为离散表示。该方法将连续、复杂的蛋白质结构空间转换为易于处理的离散格式,其码本包含4096至64000个标记,并实现高保真重建,主链均方根偏差(RMSD)约为1–5 Å。为证明所学表示的有效性,我们展示了在这些码本上训练的简单GPT模型能够生成新颖、多样且可设计的蛋白质结构。我们的方法不仅提供了蛋白质结构表示,还缓解了不同模态表示所带来的挑战,并为无缝的多模态整合奠定基础,从而增强计算方法在蛋白质设计中的能力。

英文摘要:

Representation learning and \emph{de novo} generation of proteins are pivotal computational biology tasks. Whilst natural language processing (NLP) techniques have proven highly effective for protein sequence modelling, structure modelling presents a complex challenge, primarily due to its continuous and three-dimensional nature. Motivated by this discrepancy, we introduce an approach using a vector-quantized autoencoder that effectively tokenizes protein structures into discrete representations. This method transforms the continuous, complex space of protein structures into a manageable, discrete format with a codebook ranging from 4096 to 64000 tokens, achieving high-fidelity reconstructions with backbone root mean square deviations (RMSD) of approximately 1-5 Å. To demonstrate the efficacy of our learned representations, we show that a simple GPT model trained on our codebooks can generate novel, diverse, and designable protein structures. Our approach not only provides representations of protein structure, but also mitigates the challenges of disparate modal representations and sets a foundation for seamless, multi-modal integration, enhancing the capabilities of computational methods in protein design.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑