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.

BrainGENIE: The Brain Gene Expression and Network Imputation Engine

Gene expression imputation and cell-type deconvolution in human brain with spatiotemporal precision and its implications for brain-related disorders

BRAIN-MAGNET: A novel functional genomics atlas coupled with convolutional neural networks facilitates clinical interpretation of disease relevant variants in non-coding regulatory elements

Imputed Gene Expression versus Single Nucleotide Polymorphism in Predicting Gray Matter Phenotypes.

Identification of Novel Human Glioblastoma-Specific Transcripts by Serial Analysis of Gene Expression Data Mining

NetExtractor: Extracting a Cerebellar Tissue Gene Regulatory Network Using Differentially Expressed High Mutual Information Binary RNA Profiles

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

Exploring candidate genes for human brain diseases from a brain-specific gene network.

Integration of estimated regional gene expression with neuroimaging and clinical phenotypes at biobank scale

GENet: A Graph-Based Model Leveraging Histone Marks and Transcription Factors for Enhanced Gene Expression Prediction

GENIUS: web server to predict local gene networks and key genes for biological functions

Blood-based multi-tissue gene expression inference with Bayesian ridge regression

Single-cell genomics and regulatory networks for 388 human brains

BaiHui: Cross-species Brain-specific Network Built with Hundreds of Hand-curated Datasets.

TOPMed imputed genomics enhances genomic atlas of the human proteome in brain, cerebrospinal fluid, and plasma

The Future of Genomic Profiling of Neurological Diseases Using Blood.

A transcriptomic atlas of the human brain reveals genetically determined aspects of neuropsychiatric health

Simultaneous Inference of Phenotype-Associated Genes and Relevant Tissues from GWAS Data Via Bayesian Integration of Multiple Tissue-Specific Gene Networks

Cross-tissue Graph Attention Networks for Semi-supervised Gene Expression Prediction

Brain Annotation Toolbox: Exploring the Functional and Genetic Associations of Neuroimaging Results

A Network-Based Method for Brain Disease Gene Prediction by Integrating Brain Connectome and Molecular Network.