Cis-TERN: a graph-based method to integrate single-cell chromatin accessibility with gene expression data to infer the pseudotime ordering of cells
Abstract
Motivation: Pseudotime ordering methods computationally arrange individually sequenced cells along a continuous trajectory to infer their progression through biological processes. Chromatin accessibility at regulatory elements is pivotal for orchestrating the timing of gene expression during cellular differentiation. However, data sparsity and discrete grouping imposed by clustering-based approaches, often obscures subtle cell-state transitions and intermediate populations.
Results: Here, we introduce Cis-TERN (Cis-Temporally Encoded Regulation Network), a graph-based method that integrates scRNA-seq and scATAC-seq to infer a pseudotime ordering of cells. Exploiting the fact that distal cis chromatin accessibility often precedes gene activation, Cis-TERN builds a directed bipartite graph between scRNA-seq and scATAC-seq cells with edge weights determined by either proximal or distal accessibility. Through network diffusion, Cis-TERN assigns continuous pseudotime scores to all cells. Cis-TERN infers cellular pseudotime without the need for prior biological knowledge, such as specifying a root cell type, thereby enabling unbiased reconstruction of differentiation from data alone. We show that our scores align with known lineages in healthy and tumor-derived hematopoietic cells. Additionally, Cis-TERN is able to trace individual cancer cells back to progenitor cell types. Our approach establishes a framework for leveraging chromatin accessibility and gene expression jointly to learn the timing of cell state progression.
Availability and implementation: Cis-TERN is available at https://github.com/pawelab/Cis-TERN.
Related
- Recommendation system-inspired learning of crossmodal interaction patterns between chromatin accessibility and gene expression
- PATZ1 remodels the nucleosome landscape to promote chromatin accessibility in injured neurons
- Cellular, chromatin-accessibility, and spatial contextualization of bulk transcriptional patterns linked to positive peritoneal cytology in endometrial cancer
- Genetic integration reveals causal effects of chromatin accessibility on nucleosome positioning across diverse muscle cells
- Protocol for profiling chromatin accessibility in human small-cell lung cancer FFPE-derived samples using Omni-ATAC-seq
BibTeX
@article{gokhale2026cistern,
title = {{Cis-TERN: a graph-based method to integrate single-cell chromatin accessibility with gene expression data to infer the pseudotime ordering of cells}},
author = {Ameya Gokhale and Pawel F Przytycki},
journal = {Bioinformatics Advances},
year = {2026},
month = oct,
doi = {10.1093/bioadv/vbag313},
url = {https://doi.org/10.1093/bioadv/vbag313},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Cis-TERN: a graph-based method to integrate single-cell chromatin accessibility with gene expression data to infer the pseudotime ordering of cells
AU - Ameya Gokhale
AU - Pawel F Przytycki
PY - 2026
DA - 2026/10/15
JO - Bioinformatics Advances
DO - 10.1093/bioadv/vbag313
UR - https://doi.org/10.1093/bioadv/vbag313
AB - Motivation: Pseudotime ordering methods computationally arrange individually sequenced cells along a continuous trajectory to infer their progression through biological processes. Chromatin accessibility at regulatory elements is pivotal for orchestrating the timing of gene expression during cellular differentiation. However, data sparsity and discrete grouping imposed by clustering-based approaches, often obscures subtle cell-state transitions and intermediate populations. Results: Here, we introduce Cis-TERN (Cis-Temporally Encoded Regulation Network), a graph-based method that integrates scRNA-seq and scATAC-seq to infer a pseudotime ordering of cells. Exploiting the fact that distal cis chromatin accessibility often precedes gene activation, Cis-TERN builds a directed bipartite graph between scRNA-seq and scATAC-seq cells with edge weights determined by either proximal or distal accessibility. Through network diffusion, Cis-TERN assigns continuous pseudotime scores to all cells. Cis-TERN infers cellular pseudotime without the need for prior biological knowledge, such as specifying a root cell type, thereby enabling unbiased reconstruction of differentiation from data alone. We show that our scores align with known lineages in healthy and tumor-derived hematopoietic cells. Additionally, Cis-TERN is able to trace individual cancer cells back to progenitor cell types. Our approach establishes a framework for leveraging chromatin accessibility and gene expression jointly to learn the timing of cell state progression. Availability and implementation: Cis-TERN is available at https://github.com/pawelab/Cis-TERN.
SN - 2635-0041
ER -