MMIL: A novel algorithm for disease associated cell type discovery

TL;DR

This paper introduces MMIL, a new expectation maximization algorithm that trains cell classifiers using patient-level labels, effectively identifying disease-associated cells in single-cell datasets without individual labels.

Contribution

The paper presents MMIL, a novel mixture modeling approach for multiple instance learning that enables disease cell type discovery from unlabeled single-cell data, adaptable to various classifiers.

Findings

Accurately identifies cancer cells in AML and ALL datasets.

Generalizes across tissues and treatment timepoints.

Incorporates known cell labels into training for improved performance.

Abstract

Single-cell datasets often lack individual cell labels, making it challenging to identify cells associated with disease. To address this, we introduce Mixture Modeling for Multiple Instance Learning (MMIL), an expectation maximization method that enables the training and calibration of cell-level classifiers using patient-level labels. Our approach can be used to train e.g. lasso logistic regression models, gradient boosted trees, and neural networks. When applied to clinically-annotated, primary patient samples in Acute Myeloid Leukemia (AML) and Acute Lymphoblastic Leukemia (ALL), our method accurately identifies cancer cells, generalizes across tissues and treatment timepoints, and selects biologically relevant features. In addition, MMIL is capable of incorporating cell labels into model training when they are known, providing a powerful framework for leveraging both labeled and unlabeled data simultaneously. Mixture Modeling for MIL offers a novel approach for cell classification, with significant potential to advance disease understanding and management, especially in scenarios with unknown gold-standard labels and high dimensionality.