scSTEC: Annotating Chromatin Subcompartments from Individual Single-Cell Hi-C Data via Direct Two-Dimensional Embedding and Chromosome-wise Consensus

scSTEC annotates chromatin subcompartments directly from individual single-cell Hi-C contact matrices, matching scGHOST performance across multiple datasets without imputation or cross-cell sharing.

• Why it matters: Sparse scHi-C makes subcompartment annotation difficult, and scGHOST’s deep, multi-step pipeline depends on upstream embeddings, imputation, and cross-cell information, limiting interpretability and per-cell independence.
• What they did: scSTEC uses 2D genomic embedding plus chromosome-wise consensus, leveraging scGHOST’s inter-chromosomal matrix being intrinsically rank 2 for clustering geometry, and runs without data imputation or complex preprocessing.
• The result: In bulk Hi-C, downsampled simulations, and real scHi-C data from four human cell lines, adult mouse cortex, developing mouse embryos, and developing olfactory sensory neurons, scSTEC yields comparable subcompartment annotations to scGHOST and links them to gene expression and epigenetic profiles.

AI summary of the abstract · may contain errors · not medical advice · read the paper.