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.

Scaling Dense Representations for Single Cell with Transcriptome-Scale Context

Accurate and General DNA Representations Emerge from Genome Foundation Models at Scale

CellFM: a large-scale foundation model pre-trained on transcriptomics of 100 million human cells

scGPT: toward building a foundation model for single-cell multi-omics using generative AI

Sctab: Scaling Cross-Tissue Single-Cell Annotation Models

CellMemory: Hierarchical Interpretation of Out-of-Distribution Cells Using Bottlenecked Transformer

ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy

Large-scale foundation model on single-cell transcriptomics

scGPT: Towards Building a Foundation Model for Single-Cell Multi-omics Using Generative AI

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

xTrimoGene: An Efficient and Scalable Representation Learner for Single-Cell RNA-Seq Data

Cell-Graph Compass: Modeling Single Cells with Graph Structure Foundation Model

The Development of AI Foundation Models for Single-Cell Transcriptomics

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

Transformer for one stop interpretable cell type annotation

Toward mastering the cell language by learning to generate

SCTrans: Multi-scale Scrna-Seq Sub-vector Completion Transformer for Gene-selective Cell Type Annotation

Construction of continuously expandable single-cell atlases through integration of heterogeneous datasets in a generalized cell-embedding space

HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution

Large-Scale Cell Representation Learning via Divide-and-Conquer Contrastive Learning

The Nucleotide Transformer: Building and Evaluating Robust Foundation Models for Human Genomics