Abstract:The increasing availability of large-scale single-cell datasets has enabled the detailed description of cell states across multiple biological conditions and perturbations. In parallel, recent advances in unsupervised machine learning, particularly in transfer learning, have enabled fast and scalable mapping of these new single-cell datasets onto reference atlases. The resulting large-scale machine learning models however often have millions of parameters, rendering interpretation of the newly mapped datasets challenging. Here, we propose expiMap, a deep learning model that enables interpretable reference mapping using biologically understandable entities, such as curated sets of genes and gene programs. The key concept is the substitution of the uninterpretable nodes in an autoencoder’s bottleneck by labeled nodes mapping to interpretable lists of genes, such as gene ontologies, biological pathways, or curated gene sets, for which activities are learned as constraints during reconstruction. This is enabled by the incorporation of predefined gene programs into the reference model, and at the same time allowing the model to learn de novo new programs and refine existing programs during reference mapping. We show that the model retains similar integration performance as existing methods while providing a biologically interpretable framework for understanding cellular behavior. We demonstrate the capabilities of expiMap by applying it to 15 datasets encompassing five different tissues and species. The interpretable nature of the mapping revealed unreported associations between interferon signaling via the RIG-I/MDA5 and GPCRs pathways, with differential behavior in CD8+ T cells and CD14+ monocytes in severe COVID-19, as well as the role of annexins in the cellular communications between lymphoid and myeloid compartments for explaining patient response to the applied drugs. Finally, expiMap enabled the direct comparison of a diverse set of pancreatic beta cells from multiple studies where we observed a strong, previously unreported correlation between the unfolded protein response and asparagine N-linked glycosylation. Altogether, expiMap enables the interpretable mapping of single cell transcriptome data sets across cohorts, disease states and other perturbations. ### Competing Interest Statement Fabian J. Theis consults for Immunai Inc., Singularity Bio B.V., CytoReason Ltd, and Omniscope Ltd, and has ownership interest in Dermagnostix GmbH and Cellarity.

Interpreting single-cell and spatial omics data using deep networks training dynamics

Biologically Informed Deep Learning to Infer Gene Program Activity in Single Cells

Interpretable deep learning in single-cell omics

Interpretable Fine-Grained Phenotypes of Subcellular Dynamics via Unsupervised Deep Learning

Deep Learning in Single-Cell and Spatial Transcriptomics Data Analysis: Advances and Challenges from a Data Science Perspective

A sandbox for prediction and integration of DNA, RNA, and proteins in single cells

Artificial Intelligence in Cell Annotation for High-Resolution RNA Sequencing Data

N-ACT: An Interpretable Deep Learning Model for Automatic Cell Type and Salient Gene Identification

Uncovering Interpretable Fine-Grained Phenotypes of Subcellular Dynamics through Unsupervised Self-Training of Deep Neural Networks

Transformer for one stop interpretable cell type annotation

Spatial Deconvolution of Cell Types and Cell States at Scale Utilizing TACIT

Building gene regulatory networks from scATAC-seq and scRNA-seq using Linked Self Organizing Maps

SORBET: Automated cell-neighborhood analysis of spatial transcriptomics or proteomics for interpretable sample classification via GNN

Deep Neural Networks to Register and Annotate the Cells of the C. elegans Nervous System

JIND: joint integration and discrimination for automated single-cell annotation

Learning cell-specific networks from dynamics and geometry of single cells

A gene regulatory network-aware graph learning method for cell identity annotation in single-cell RNA-seq data

Deep learning exploration of single-cell and spatially resolved cancer transcriptomics to unravel tumour heterogeneity

Mapping single-cell developmental potential in health and disease with interpretable deep learning

A self-training interpretable cell type annotation framework using specific marker gene

TACCO unifies annotation transfer and decomposition of cell identities for single-cell and spatial omics