Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 38 min ago
Peer-reviewedCC BY 4.0Communications Biology

OMIDIENT: Multiomics Integration for Cancer by Dirichlet Auto-Encoder Networks

  • Negar Safinianaini1
  • Niko Välimäki2,3
  • Roman Bresson4
  • Alexandra Gorbonos1
  • Kristiina Rajamäki2,3
  • Lauri A. Aaltonen2,3,5
  • Pekka Marttinen1
5 affiliations
  1. 1Dept. of Computer Science, Aalto University, Helsinki, Finland
  2. 2Dept. of Medical and Clinical Genetics, University of Helsinki, Helsinki, Finland
  3. 3Applied Tumor Genomics Research Program, Research Programs Unit, University of Helsinki, Helsinki, Finland
  4. 4Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, UAE
  5. 5iCAN Digital Precision Cancer Medicine Flagship, University of Helsinki, Helsinki, Finland
Read full textPDFDOICite

Abstract

To achieve a more comprehensive understanding of cancer, novel computational methods are required for the integrative analysis of data from different molecular layers, such as genomics, transcriptomics, and epigenomics. Here, we present an innovative multi-omics integrative method that performs unsupervised representation learning, referred to as OMIDIENT: multiOMics Integration for cancer by DIrichlet auto-ENcoder neTworks. OMIDIENT provides a natural framework for modeling sparse and compositional latent representations by employing a deep generative model, where the latent space is distributed as the product of Dirichlet distributions. Applied to five different cancers, we demonstrate that OMIDIENT outperforms the top state-of-the-art unsupervised multi-omics integrative analysis approaches in clustering, classification, and reconstruction of missing data using mRNA expression data, DNA methylation data, and microRNA expression data. Furthermore, we provide interpretability analyses for OMIDIENT that not only support its improved performance, but also offer valuable insights into the underlying structure captured by the learned representations.

Related

  1. Multi-Omic Integration of the miRNome and Transcriptome in Dermal Fibroblasts from Hypermobile Ehlers–Danlos Syndrome and Hypermobility Spectrum Disorders Identifies Convergent Candidate miRNA-mRNA Regulatory Networks Linking Extracellular Matrix Remodeling and Inflammation
    Iulia Andreea Pelisenco, Valeria Cinquina, Nicola Chiarelli +10
    International journal of molecular sciencesPeer-reviewedCC BYFull textGeneral biochemistryEurope PMC
  2. Decomposing multi-scale dynamic regulation from single-cell multiomics with scMagnify
    Xufeng Chen, Xi Yan, Bihan Shen +7
    Nature CommunicationsPeer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationCrossref
  3. A generalizable Hi-C foundation model for chromatin architecture, single-cell and multiomics analysis across species
    Xiao Wang, Yuanyuan Zhang, Suhita Ray +6
    Nature MethodsPeer-reviewedLicence not stated9 citedMolecular biology & gene regulationOpenAlex
  4. Multiomics analysis reveals the enhancer landscapes and 3D chromatin architecture in esophageal squamous cell carcinoma
    Haoqiang Sun, Yongjie Wang, Wenwen Xu +9
    Communications BiologyPeer-reviewedCC BY-NC-ND 4.0Full textMolecular biology & gene regulationCrossref
  5. Roadmap for inferring metabolic regulation using multiomics, machine learning, and genome-scale modeling.
    Ryan Schildcrout, Nicole Rhoads, Sriram Chandrasekaran
    Npj metabolic health and diseasePeer-reviewedCC BY-NC-NDFull textMolecular biology & gene regulationEurope PMCPMC

Cite

BibTeXRIS
BibTeX
@article{safinianaini2026omidient,
  title = {{OMIDIENT: Multiomics Integration for Cancer by Dirichlet Auto-Encoder Networks}},
  author = {Negar Safinianaini and Niko Välimäki and Roman Bresson and Alexandra Gorbonos and Kristiina Rajamäki and Lauri A. Aaltonen and Pekka Marttinen},
  journal = {Communications Biology},
  year = {2026},
  month = jul,
  doi = {10.1038/s42003-026-10578-1},
  url = {https://doi.org/10.1038/s42003-026-10578-1},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - OMIDIENT: Multiomics Integration for Cancer by Dirichlet Auto-Encoder Networks
AU  - Negar Safinianaini
AU  - Niko Välimäki
AU  - Roman Bresson
AU  - Alexandra Gorbonos
AU  - Kristiina Rajamäki
AU  - Lauri A. Aaltonen
AU  - Pekka Marttinen
PY  - 2026
DA  - 2026/07/06
JO  - Communications Biology
DO  - 10.1038/s42003-026-10578-1
UR  - https://doi.org/10.1038/s42003-026-10578-1
AB  - To achieve a more comprehensive understanding of cancer, novel computational methods are required for the integrative analysis of data from different molecular layers, such as genomics, transcriptomics, and epigenomics. Here, we present an innovative multi-omics integrative method that performs unsupervised representation learning, referred to as OMIDIENT: multiOMics Integration for cancer by DIrichlet auto-ENcoder neTworks. OMIDIENT provides a natural framework for modeling sparse and compositional latent representations by employing a deep generative model, where the latent space is distributed as the product of Dirichlet distributions. Applied to five different cancers, we demonstrate that OMIDIENT outperforms the top state-of-the-art unsupervised multi-omics integrative analysis approaches in clustering, classification, and reconstruction of missing data using mRNA expression data, DNA methylation data, and microRNA expression data. Furthermore, we provide interpretability analyses for OMIDIENT that not only support its improved performance, but also offer valuable insights into the underlying structure captured by the learned representations.
SN  - 2399-3642
ER  -