Transformer-based Single-Cell Language Model: A Survey
Wei Lan, Guohang He, Mingyang Liu, Qingfeng Chen, Junyue Cao, Wei Peng
TL;DR
This survey examines how Transformer architectures are applied to single-cell omics, covering both language-modeling and large-language-model approaches across scRNA-seq, scATAC-seq, spatial omics, and multi-omics. It catalogs representative models, architectures, and training strategies, and maps them to downstream tasks such as batch correction, clustering, annotation, GRN inference, and perturbation prediction. Key contributions include a structured taxonomy of single-cell Transformers, critical evaluation of downstream performance, and a discussion of challenges like long sequence handling, overfitting, computation, and interpretability. The work highlights practical implications for leveraging Transformers in single-cell biology and outlines directions for more scalable, interpretable, and generalizable solutions.
Abstract
The transformers have achieved significant accomplishments in the natural language processing as its outstanding parallel processing capabilities and highly flexible attention mechanism. In addition, increasing studies based on transformers have been proposed to model single-cell data. In this review, we attempt to systematically summarize the single-cell language models and applications based on transformers. First, we provide a detailed introduction about the structure and principles of transformers. Then, we review the single-cell language models and large language models for single-cell data analysis. Moreover, we explore the datasets and applications of single-cell language models in downstream tasks such as batch correction, cell clustering, cell type annotation, gene regulatory network inference and perturbation response. Further, we discuss the challenges of single-cell language models and provide promising research directions. We hope this review will serve as an up-to-date reference for researchers interested in the direction of single-cell language models.