scMILD: Single-cell multiple instance learning for sample classification and associated subpopulation discovery

TL;DR

scMILD is a machine learning framework that identifies cell subpopulations linked to diseases using only sample-level labels.

Contribution

scMILD introduces a weakly supervised multiple instance learning framework for subpopulation discovery without requiring cell-level labels.

Findings

scMILD successfully identifies condition-associated cells in single-cell datasets using only sample-level labels.

The method reveals monocyte state transitions in COVID-19 progression and distinguishes shared and disease-specific signatures in Lupus and COVID-19.

Validation on diverse disease datasets confirms scMILD's ability to retrieve known biological signatures.

Abstract

Linking cellular states to clinical phenotypes is a major challenge in single-cell analysis. Here, we present single-cell multiple instance learning for sample classification and associated subpopulation discovery (scMILD), a weakly supervised multiple instance learning framework that robustly identifies condition-associated cells using only sample-level labels. After systematically validating scMILD’s accuracy through controlled simulations, we applied it to diverse disease datasets, confirming its ability to retrieve known biological signatures. Building on this, our sample-informed analysis of scMILD-identified monocytes in COVID-19 revealed a temporal transition from an early antiviral to a late stress-response state. Furthermore, in a cross-disease application, a model trained on COVID-19 successfully stratified patients with Lupus and distinguished shared inflammatory states from disease-specific ones. scMILD thus provides a validated and versatile strategy to dissect cellular heterogeneity, bridging single-cell observations with high-level phenotypes.

•scMILD identifies condition-associated cells using only sample-level labels•Dual-branch model with orthogonal projection loss improves cell separation•Sample-informed analysis reveals monocyte state transitions in COVID-19 progression•Cross-disease modeling reveals shared and specific signatures in SLE and COVID-19

scMILD identifies condition-associated cells using only sample-level labels

Dual-branch model with orthogonal projection loss improves cell separation

Sample-informed analysis reveals monocyte state transitions in COVID-19 progression

Cross-disease modeling reveals shared and specific signatures in SLE and COVID-19

Biocomputational method; Classification of bioinformatical subject; Transcriptomics

Figures