Robust Multi-view Co-expression Network Inference

Teodora Pandeva, Martijs Johannes Jonker, Leendert Hamoen, Joris Mooij, Patrick Forré
Proceedings of the Fourth Conference on Causal Learning and Reasoning, PMLR 275:490-513, 2025.

Abstract

Unraveling the co-expression of genes across studies enhances the understanding of cellular processes. Inferring gene co-expression networks from transcriptome data presents many challenges, including the high-dimensionality of the data relative to the number of samples, sample correlations, and batch effects. To address these complexities, we introduce a robust method for high-dimensional graph inference from multiple independent studies. We base our approach on the premise that each dataset is essentially a noisy linear mixture of gene loadings that follow a multivariate $t$-distribution with a sparse precision matrix, which is shared across studies. This allows us to show that we can identify the co-expression matrix up to a scaling factor among other model parameters. Our method employs an Expectation-Maximization procedure for parameter estimation. Empirical evaluation on synthetic and gene expression data demonstrates our method’s improved ability to learn the underlying graph structure compared to baseline methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v275-pandeva25a, title = {Robust Multi-view Co-expression Network Inference}, author = {Pandeva, Teodora and Jonker, Martijs Johannes and Hamoen, Leendert and Mooij, Joris and Forr\'{e}, Patrick}, booktitle = {Proceedings of the Fourth Conference on Causal Learning and Reasoning}, pages = {490--513}, year = {2025}, editor = {Huang, Biwei and Drton, Mathias}, volume = {275}, series = {Proceedings of Machine Learning Research}, month = {07--09 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v275/main/assets/pandeva25a/pandeva25a.pdf}, url = {https://proceedings.mlr.press/v275/pandeva25a.html}, abstract = {Unraveling the co-expression of genes across studies enhances the understanding of cellular processes. Inferring gene co-expression networks from transcriptome data presents many challenges, including the high-dimensionality of the data relative to the number of samples, sample correlations, and batch effects. To address these complexities, we introduce a robust method for high-dimensional graph inference from multiple independent studies. We base our approach on the premise that each dataset is essentially a noisy linear mixture of gene loadings that follow a multivariate $t$-distribution with a sparse precision matrix, which is shared across studies. This allows us to show that we can identify the co-expression matrix up to a scaling factor among other model parameters. Our method employs an Expectation-Maximization procedure for parameter estimation. Empirical evaluation on synthetic and gene expression data demonstrates our method’s improved ability to learn the underlying graph structure compared to baseline methods.} }
Endnote
%0 Conference Paper %T Robust Multi-view Co-expression Network Inference %A Teodora Pandeva %A Martijs Johannes Jonker %A Leendert Hamoen %A Joris Mooij %A Patrick Forré %B Proceedings of the Fourth Conference on Causal Learning and Reasoning %C Proceedings of Machine Learning Research %D 2025 %E Biwei Huang %E Mathias Drton %F pmlr-v275-pandeva25a %I PMLR %P 490--513 %U https://proceedings.mlr.press/v275/pandeva25a.html %V 275 %X Unraveling the co-expression of genes across studies enhances the understanding of cellular processes. Inferring gene co-expression networks from transcriptome data presents many challenges, including the high-dimensionality of the data relative to the number of samples, sample correlations, and batch effects. To address these complexities, we introduce a robust method for high-dimensional graph inference from multiple independent studies. We base our approach on the premise that each dataset is essentially a noisy linear mixture of gene loadings that follow a multivariate $t$-distribution with a sparse precision matrix, which is shared across studies. This allows us to show that we can identify the co-expression matrix up to a scaling factor among other model parameters. Our method employs an Expectation-Maximization procedure for parameter estimation. Empirical evaluation on synthetic and gene expression data demonstrates our method’s improved ability to learn the underlying graph structure compared to baseline methods.
APA
Pandeva, T., Jonker, M.J., Hamoen, L., Mooij, J. & Forré, P.. (2025). Robust Multi-view Co-expression Network Inference. Proceedings of the Fourth Conference on Causal Learning and Reasoning, in Proceedings of Machine Learning Research 275:490-513 Available from https://proceedings.mlr.press/v275/pandeva25a.html.

Related Material