Single-cell omics technologies have enabled the analysis of individual cells within a biological sample, leading to a more detailed understanding of biological systems. An essential goal in single-cell RNA-seq (scRNA-seq) analysis is accurately determining the cell type of each cell. However, existing single-cell annotation methods face the challenge of effectively processing large-scale datasets. Additionally, with the increasing availability of scRNA-seq datasets, integrating multiple datasets and addressing batch effects become challenges in cell-type annotation.

To tackle these challenges, the Plant Bioinformatics Group and the Molecular Breeding of Turfgrass and Forage Grass Group at Wuhan Botanical Garden, have developed a supervised method called CIForm (cell-type identification based on Transformer), which has been successfully used for Generative Pre-trained Transformer, such as chatGPT.

To evaluate CIForm's effectiveness and robustness, researchers compared it with leading tools using benchmark datasets. Through systematic comparisons under various cell-type annotation scenarios, it demonstrates that CIForm excels in cell-type annotation.

The research was conducted by XU Jing under the supervision of Prof. ZHANG Xiujun and Prof. CHEN Liang. This research was partially supported by the National Natural Science Foundation of China.

Results have been published in Briefings in Bioinformatics, entitled "CIForm as a Transformer-based model for cell-type annotation of large-scale single-cell RNA-seq data".

Code Link: https://github.com/zhanglab-wbgcas/CIForm