Answers are grounded in this paper and cite their sources.
Single-cell sequencing provides transcriptomic profiling at single-cell resolution, uncovering cellular heterogeneity with unprecedented precision. Yet, current single cell data analysis suffers from the inherent data noises, batch effects, and sparsity, highlighting the requirement of a unified model to represent cellular states. To circumvent this problem, many recent efforts focus on training single-cell foundation models based on large datasets. However, current human foundation models are still limited by the sizes of training data and model parameters. Here, we have collected a diverse dataset of 100 million human cells, on which we train a single-cell foundation model (CellFM) containing 800 million parameters. To balance efficiency and performance, the model is trained through a modified RetNet framework on the MindSpore. Extensive experiments have shown that CellFM outperforms existing models in cell annotation, perturbation prediction, gene function prediction, and gene-gene relationship capturing. Single-cell sequencing reveals cellular heterogeneity but is challenged by technical noise and batch effects. Here, authors present CellFM, an 800-million-parameter foundation model trained on 100 million human cells through the MindSpore framework, which outperforms existing models in downstream tasks.
No takes yet. Share an insight, caveat, or question.
Zeng et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: