A Cross-Species Generative Cell Atlas Across 1.5 Billion Years of Evolution: The TranscriptFormer Single-cell Model
Preprint
Abstract
Single-cell transcriptomics has revolutionized our understanding of cellular diversity, yet our understanding of the transcriptional programs across the tree of life remains limited. Here we present TranscriptFormer, a family of generative foundation models trained on up to 112 million cells spanning 1.53 billion years of evolution across 12 species. By jointly modeling gene identities and expression levels using a novel generative architecture, TranscriptFormer encodes multi-scale biological structure, functioning as a queryable virtual cell atlas. We demonstrate state-of-the-art performance on both in-distribution and out-of-distribution cell type classification, with robust performance even for species separated by over 685 million years of evolution. TranscriptFormer can also perform zero-shot disease state identification in human cells and accurately transfers cell state annotations across species boundaries. As a generative model, TranscriptFormer can be prompted to predict cell type-specific transcription factors and gene-gene interactions that align with independent experimental observations. Developmental trajectories, phylogenetic relationships and cellular hierarchies emerge naturally in TranscriptFormer’s representations without any explicit training on these annotations. This work establishes a powerful framework for quantitative single-cell analysis, and comparative cellular biology, thus demonstrating that universal principles of cellular organization can be learned and predicted across the tree of life.Keywords
This publication has 25 references indexed in Scilit:
- CZ CELLxGENE Discover: a single-cell data platform for scalable exploration, analysis and modeling of aggregated dataNucleic Acids Research, 2024
- scMulan: A Multitask Generative Pre-Trained Language Model for Single-Cell Analysis2024
- Interstitial macrophages are a focus of viral takeover and inflammation in COVID-19 initiation in human lungJournal of Experimental Medicine, 2024
- Embryo-scale reverse genetics at single-cell resolutionNature, 2023
- An atlas of rabbit development as a model for single-cell comparative genomicsNature Cell Biology, 2023
- Integrated analysis of multimodal single-cell dataCell, 2021
- Comprehensive Integration of Single-Cell DataCell, 2019
- The single-cell transcriptional landscape of mammalian organogenesisNature, 2019
- Deep generative modeling for single-cell transcriptomicsNature Methods, 2018
- Integrating single-cell transcriptomic data across different conditions, technologies, and speciesNature Biotechnology, 2018