Learning multi-cellular representations of single-cell transcriptomics data enables characterization of patient-level disease states
Clinical Snapshot
PICO Framework
| P — Population | Over 2,700 patients with scRNA-seq data from 12.5 million cells across various disease states |
| I — Intervention | PaSCient machine learning model for multi-cellular representation learning |
| C — Comparator | Existing single-cell foundation models and traditional analysis methods |
| O — Outcomes | Disease classification accuracy, dimensionality reduction performance, feature prioritization, and patient subgroup discovery |
Bottom Line
PaSCient represents a significant methodological advance in single-cell genomics by generating patient-level disease representations from multi-cellular transcriptomic contexts. The approach addresses a critical gap in current single-cell analysis methods, which typically focus on individual cell types rather than tissue ecosystems. With validation on 12.5 million cells from over 2,700 patients, the method demonstrates superior performance in disease classification and provides interpretable results through cell and gene-level importance scores. This could enhance precision medicine approaches by enabling better patient stratification and disease subtyping. However, the abstract lacks specific performance metrics and statistical validation details. Clinical implementation would require robust computational infrastructure and high-quality scRNA-seq data. The method's ability to identify patient subgroups and prioritize disease-relevant features makes it particularly valuable for complex diseases where cellular interactions drive pathology. Further validation in independent cohorts and assessment of clinical utility will be essential for translation to routine practice.
Key Findings
P Value: Not reported
Effect Size: Not specified in abstract
Primary Outcome: Superior performance in disease classification compared to existing methods
Nnt Or Sensitivity: Classification accuracy and feature importance scores provided but specific values not reported in abstract
Confidence Interval: Not reported
Clinical Application
Requires access to high-quality scRNA-seq data and computational infrastructure for implementation Relevant for Australian precision medicine initiatives and research institutions with genomics capabilities. Could support TGA regulatory submissions requiring transcriptomic evidence and align with NHMRC precision medicine priorities Patients with diseases amenable to single-cell transcriptomic analysis, particularly complex multi-cellular conditions
Abstract
Single-cell RNA sequencing (scRNA-seq) is a prominent tool for studying human disease biology. The availability of massive scRNA-seq datasets and advanced machine learning has driven the development of single-cell foundation models that provide versatile cell representations based on expression. To understand disease states, we need to account for entire tissue ecosystems, simultaneously considering many different interacting cells. Here, we tackle this challenge by generating patient-level representations derived from multi-cellular expression contexts measured with scRNA-seq of tissues. We develop PaSCient, a machine learning model that employs a multi-level representation learning paradigm and provides importance scores at the cell and gene levels for fine-grained analysis of disease characteristics. We apply PaSCient to a large-scale scRNA-seq atlas of 12.5 million cells from over 2,700 patients. Comprehensive benchmarking demonstrates the superiority of PaSCient in disease classification and multiple downstream applications, including dimensionality reduction, feature prioritization, and patient subgroup discovery.
References
- 1.Liu, T., De Brouwer, E., Verma, A., Missarova, A., Kuo, T., Diamant, N., Wang, H., Hao, M., Biancalani, T., Bravo, H. C., Scalia, G., Regev, A., & Heimberg, G. (2026). Learning multi-cellular representations of single-cell transcriptomics data enables characterization of patient-level disease states. Cell Systems. https://doi.org/10.1016/j.cels.2026.101570
Related Research
Current opinion in chemical biology
Prediction of protein-protein interactions and co-complex models with deep learning
3 Aug 2026
Journal of chemical information and modeling
ProphDR: An Interpretable Deep Learning Model for Predicting Cancer Drug Response via Multi-Omics and Cross-Attention Mechanisms
28 July 2026
PloS one
Multi-view graph-regularized deep metric subspace clustering network
26 July 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service