Abstract:Single-cell Assay for Transposase-Accessible Chromatin with sequencing (scATAC-seq) has emerged as a powerful technique for investigating open chromatin landscapes at single-cell resolution. However, scATAC-seq cell representation learning and its downstream tasks remain challenging due to the high dimensional, sparse, and noisy properties of the data. Genome Foundation Models (GFMs), which are pre-trained on millions of DNA sequences via masked nucleotide prediction, have proven effective at analyzing genomic sequences. Given that open chromatin regions often harbour salient sequence features and that the application of GFMs to single-cell biology is underexplored, we hypothesize that leveraging GFMs' sequence embeddings can improve the accuracy and generalizability of scATAC-seq modeling. In this study, we introduce the Genome Foundation Embedded Topic Model (GFETM), an interpretable and transferable deep learning framework that combines GFMs with the Embedded Topic Model (ETM) for scATAC-seq data analysis. We show that by probing and integrating the DNA sequence embeddings extracted by a GFM from open chromatin regions (OCRs), GFETM not only achieves state-of-the-art performance in scATAC-seq cell representation learning and peak imputation on scATAC-seq datasets at scale but also demonstrates generalizability and transferability to single-cell transcriptomes and across different subjects, tissues, and species. Furthermore, GFETM is capable of improving the detection of marker peaks associated with known cell-type specific marker genes in human hematopoietic stem cells (HSCs). Finally, the latent topic mixtures inferred by GFETM reveal biologically meaningful epigenomic signatures that are indicative of kidney diabetes as well as sequence motif signatures for human HSC differentiation.

RefTM: reference-guided topic modeling of single-cell chromatin accessibility data

Scart: Recognizing Cell Clusters and Constructing Trajectory from Single-Cell Epigenomic Data

RA3 is a reference-guided approach for epigenetic characterization of single cells

scATAC-Ref: a reference of scATAC-seq with known cell labels in multiple species

simCAS: an embedding-based method for simulating single-cell chromatin accessibility sequencing data

GFETM: Genome Foundation-based Embedded Topic Model for scATAC-seq Modeling

scCASE: accurate and interpretable enhancement for single-cell chromatin accessibility sequencing data

Information retrieval in single cell chromatin analysis using TF-IDF transformation methods

scAMACE: Model-based approach to the joint analysis of single-cell data on chromatin accessibility, gene expression and methylation

CAraCAl: CAMML with the integration of chromatin accessibility

Scican: Single-cell Chromatin Accessibility and Gene Expression Data Integration Via Cycle-consistent Adversarial Network

Cofea: Correlation-Based Feature Selection for Single-Cell Chromatin Accessibility Data

Fundamental and Practical Approaches for Single-Cell ATAC-seq Analysis

scCTS: identifying the cell type-specific marker genes from population-level single-cell RNA-seq

Characterizing cellular heterogeneity in chromatin state with scCUT&Tag-pro

MOCHA's advanced statistical modeling of scATAC-seq data enables functional genomic inference in large human cohorts

Reference panel-guided super-resolution inference of Hi-C data

scTIM: seeking cell-type-indicative marker from single cell RNA-seq data by consensus optimization

A Biophysical Model for ATAC-seq Data Analysis

Global prediction of chromatin accessibility using small-cell-number and single-cell RNA-seq

Profiling the genetic determinants of chromatin accessibility with scalable single-cell CRISPR screens