Understanding Transcriptional Regulatory Redundancy by Learnable Global Subset Perturbations

Junhao Liu, Siwei Xu, Dylan Riffle, Ziheng Duan, Martin Renqiang Min, Jing Zhang
Proceedings of the 16th Asian Conference on Machine Learning, PMLR 260:383-398, 2025.

Abstract

Transcriptional regulation through cis-regulatory elements (CREs) is crucial for numerous biological functions, with its disruption potentially leading to various diseases. These CREs often exhibit redundancy, allowing them to compensate for each other in response to external disturbances, highlighting the need for methods to identify CRE sets that collaboratively regulate gene expression effectively. To address this, we introduce GRIDS, an in silico computational method that approaches the task as a global feature explanation challenge to dissect combinatorial CRE effects in two phases. First, GRIDS constructs a differentiable surrogate function to mirror the complex gene regulatory process, facilitating cross-translations in single-cell modalities. It then employs learnable perturbations within a state transition framework to offer global explanations, efficiently navigating the combinatorial feature landscape. Through comprehensive benchmarks, GRIDS demonstrates superior explanatory capabilities compared to other leading methods. Moreover, GRIDS’s global explanations reveal intricate regulatory redundancy across cell types and states, underscoring its potential to advance our understanding of cellular regulation in biological research.

Cite this Paper

BibTeX
@InProceedings{pmlr-v260-liu25a, title = {Understanding Transcriptional Regulatory Redundancy by Learnable Global Subset Perturbations}, author = {Liu, Junhao and Xu, Siwei and Riffle, Dylan and Duan, Ziheng and Min, Martin Renqiang and Zhang, Jing}, booktitle = {Proceedings of the 16th Asian Conference on Machine Learning}, pages = {383--398}, year = {2025}, editor = {Nguyen, Vu and Lin, Hsuan-Tien}, volume = {260}, series = {Proceedings of Machine Learning Research}, month = {05--08 Dec}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v260/main/assets/liu25a/liu25a.pdf}, url = {https://proceedings.mlr.press/v260/liu25a.html}, abstract = {Transcriptional regulation through cis-regulatory elements (CREs) is crucial for numerous biological functions, with its disruption potentially leading to various diseases. These CREs often exhibit redundancy, allowing them to compensate for each other in response to external disturbances, highlighting the need for methods to identify CRE sets that collaboratively regulate gene expression effectively. To address this, we introduce GRIDS, an in silico computational method that approaches the task as a global feature explanation challenge to dissect combinatorial CRE effects in two phases. First, GRIDS constructs a differentiable surrogate function to mirror the complex gene regulatory process, facilitating cross-translations in single-cell modalities. It then employs learnable perturbations within a state transition framework to offer global explanations, efficiently navigating the combinatorial feature landscape. Through comprehensive benchmarks, GRIDS demonstrates superior explanatory capabilities compared to other leading methods. Moreover, GRIDS’s global explanations reveal intricate regulatory redundancy across cell types and states, underscoring its potential to advance our understanding of cellular regulation in biological research.} }
Endnote
%0 Conference Paper %T Understanding Transcriptional Regulatory Redundancy by Learnable Global Subset Perturbations %A Junhao Liu %A Siwei Xu %A Dylan Riffle %A Ziheng Duan %A Martin Renqiang Min %A Jing Zhang %B Proceedings of the 16th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Vu Nguyen %E Hsuan-Tien Lin %F pmlr-v260-liu25a %I PMLR %P 383--398 %U https://proceedings.mlr.press/v260/liu25a.html %V 260 %X Transcriptional regulation through cis-regulatory elements (CREs) is crucial for numerous biological functions, with its disruption potentially leading to various diseases. These CREs often exhibit redundancy, allowing them to compensate for each other in response to external disturbances, highlighting the need for methods to identify CRE sets that collaboratively regulate gene expression effectively. To address this, we introduce GRIDS, an in silico computational method that approaches the task as a global feature explanation challenge to dissect combinatorial CRE effects in two phases. First, GRIDS constructs a differentiable surrogate function to mirror the complex gene regulatory process, facilitating cross-translations in single-cell modalities. It then employs learnable perturbations within a state transition framework to offer global explanations, efficiently navigating the combinatorial feature landscape. Through comprehensive benchmarks, GRIDS demonstrates superior explanatory capabilities compared to other leading methods. Moreover, GRIDS’s global explanations reveal intricate regulatory redundancy across cell types and states, underscoring its potential to advance our understanding of cellular regulation in biological research.
APA
Liu, J., Xu, S., Riffle, D., Duan, Z., Min, M.R. & Zhang, J.. (2025). Understanding Transcriptional Regulatory Redundancy by Learnable Global Subset Perturbations. Proceedings of the 16th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 260:383-398 Available from https://proceedings.mlr.press/v260/liu25a.html.

Related Material