加拿大大学健康网络Bo Wang研究组近日取得一项新成果。他们提出了scGPT——利用人工智能构建单细胞多组学基础模型。相关论文于2024年2月26日发表在《自然-方法学》杂志上。

研究人员将语言和细胞生物学相类比，探讨了基础模型在推进细胞生物学和基因研究方面的适用性。研究利用快速扩增的单细胞测序数据，构建了一个单细胞生物学基础模型-scGPT，该模型基于一个生成式预训练变换器，包涵超过3300万个细胞的存储库。

研究结果表明，scGPT能有效地提炼出有关基因和细胞的关键生物学特点。通过对迁移学习的进一步调整，scGPT得以优化，从而在各种下游应用中展现出卓越的性能。这包括细胞类型注释、多批次整合、多原子整合、扰动反应预测和基因网络推断等功能。

据悉，生成式预训练模型在语言和计算机视觉等多个领域都取得了显著成功。具体来说，大规模多样化数据集与预训练转换器的结合，已成为研发基础模型的一种潜在方法。

附：英文原文

Title: scGPT: toward building a foundation model for single-cell multi-omics using generative AI

Author: Cui, Haotian, Wang, Chloe, Maan, Hassaan, Pang, Kuan, Luo, Fengning, Duan, Nan, Wang, Bo

Issue&Volume: 2024-02-26

Abstract: Generative pretrained models have achieved remarkable success in various domains such as language and computer vision. Specifically, the combination of large-scale diverse datasets and pretrained transformers has emerged as a promising approach for developing foundation models. Drawing parallels between language and cellular biology (in which texts comprise words; similarly, cells are defined by genes), our study probes the applicability of foundation models to advance cellular biology and genetic research. Using burgeoning single-cell sequencing data, we have constructed a foundation model for single-cell biology, scGPT, based on a generative pretrained transformer across a repository of over 33 million cells. Our findings illustrate that scGPT effectively distills critical biological insights concerning genes and cells. Through further adaptation of transfer learning, scGPT can be optimized to achieve superior performance across diverse downstream applications. This includes tasks such as cell type annotation, multi-batch integration, multi-omic integration, perturbation response prediction and gene network inference.

DOI: 10.1038/s41592-024-02201-0

Source: https://www.nature.com/articles/s41592-024-02201-0