IST: an ontology-guided attention-based autoencoder for interpretable analysis of single-cell transcriptomic data
Open Access
- 24 April 2026
- journal article
- Published by Springer Nature in Scientific Reports
- Vol. 16 (1)
Abstract
Extracting biologically interpretable insights from single-cell RNA sequencing (scRNA-seq) data remains a major challenge. While deep learning approaches have demonstrated strong predictive performance for dimensionality reduction and representation learning, their latent representations often lack clear biological meaning, and prior biological knowledge is frequently incorporated only weakly or inconsistently. As a result, learned features may not faithfully reflect known gene-function relationships, limiting biological interpretability and downstream insight. We present IST (Interpretation of Single-cell Transcriptomic data), a novel attention-based framework designed to reliably integrate biological prior knowledge into the analysis of single-cell transcriptomic data. IST incorporates gene ontology structure directly into the model architecture and employs attention mechanisms to capture context-dependent gene activity. A tailored correlation-based loss enforces alignment between learned gene activities and known gene-function relationships, ensuring that biological priors are faithfully reflected in the learned representations rather than acting as superficial regularization. We evaluate IST on three scRNA-seq datasets. Across all datasets, IST identifies condition-specific gene activity patterns, uncovers novel relationships among biological processes, and infers previously unannotated gene functions supported by existing literature. Compared with state-of-the-art methods, IST provides more interpretable and biologically grounded representations of gene functions and activities, while maintaining competitive predictive performance. The datasets and tool are available at 10.6084/m9.figshare.30811334.Keywords
Funding Information
- National Science Foundation (2514869)
- National Science Foundation (2120907)
This publication has 52 references indexed in Scilit:
- scSwinFormer: A Transformer-Based Cell-Type Annotation Method for scRNA-Seq Data Using Smooth Gene Embedding and Global FeaturesJournal of Chemical Information and Modeling, 2024
- New insights into the pathogenicity of TMEM165 variants using structural modeling based on AlphaFold 2 predictionsComputational and Structural Biotechnology Journal, 2023
- INSISTC: Incorporating network structure information for single-cell type classificationGenomics, 2022
- An injury-responsive Rac-to-Rho GTPase switch drives activation of muscle stem cells through rapid cytoskeletal remodelingCell Stem Cell, 2022
- Key concepts in muscle regeneration: muscle “cellular ecology” integrates a gestalt of cellular cross-talk, motility, and activity to remodel structure and restore functionEuropean Journal of Applied Physiology, 2021
- Predicting Drug Response and Synergy Using a Deep Learning Model of Human Cancer CellsCancer Cell, 2020
- Macrophage-derived Wnt signaling increases endothelial permeability during skeletal muscle injuryInflammation Research, 2020
- Single-Cell Analysis of the Muscle Stem Cell Hierarchy Identifies Heterotypic Communication Signals Involved in Skeletal Muscle RegenerationCell Reports, 2020
- TMEM165 deficiencies in Congenital Disorders of Glycosylation type II (CDG-II): Clues and evidences for roles of the protein in Golgi functions and ion homeostasisTissue and Cell, 2017
- PECAM-1: Conflicts of interest in inflammationLife Sciences, 2010