Omics BioAnalytics : an RShiny application for multimodal biomarker panel discovery and assessment
(2026) In Bioinformatics Advances 6(1).- Abstract
Motivation Machine learning offers a powerful approach for building predictive models from high-dimensional molecular data. Omics technologies such as transcriptomics, proteomics, and metabolomics quantify thousands of molecules simultaneously, providing deep insights into disease biology. Integrating multiple modalities can enhance predictive performance, as shown in histology-omics and holter-omics applications. To support streamlined, reproducible, and user-friendly multimodal analytics, we developed Omics BioAnalytics, an R Shiny platform for unified analysis, integration, and interpretation of diverse omics datasets. Results Omics BioAnalytics performs late integration using ensembles of elastic net models trained independently on... (More)
Motivation Machine learning offers a powerful approach for building predictive models from high-dimensional molecular data. Omics technologies such as transcriptomics, proteomics, and metabolomics quantify thousands of molecules simultaneously, providing deep insights into disease biology. Integrating multiple modalities can enhance predictive performance, as shown in histology-omics and holter-omics applications. To support streamlined, reproducible, and user-friendly multimodal analytics, we developed Omics BioAnalytics, an R Shiny platform for unified analysis, integration, and interpretation of diverse omics datasets. Results Omics BioAnalytics performs late integration using ensembles of elastic net models trained independently on each modality, with predictions averaged across datasets. The platform provides interactive dashboards for metadata exploration, exploratory analyses, differential expression, gene set analysis, and biomarker discovery. Results are visualized through dynamic plots and downloadable reports, ensuring transparent and reproducible workflows. A unique feature is the integrated multimodal Alexa Skill, which enables voice-based querying and rapid visualization. Together, these web and voice-enabled tools offer accessible and reproducible multimodal analytics for biomedical researchers, supporting the discovery of molecular signatures, predictive biomarkers, and therapeutic targets. Availability and implementation All source code, public datasets, video walkthroughs, and the deployed application are available at: https://github.com/CompBio-Lab/omicsBioAnalytics.
(Less)
- author
- Dyce, Josh ; Rieskamp, Lea LU ; Tebbutt, Scott J. ; McManus, Bruce M. and Singh, Amrit
- organization
- publishing date
- 2026
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Bioinformatics Advances
- volume
- 6
- issue
- 1
- article number
- vbaf307
- publisher
- Oxford University Press
- external identifiers
-
- scopus:105027167952
- pmid:41523655
- ISSN
- 2635-0041
- DOI
- 10.1093/bioadv/vbaf307
- language
- English
- LU publication?
- yes
- id
- a6b3488b-14d5-48f3-b503-7b3782eb8708
- date added to LUP
- 2026-03-25 11:40:28
- date last changed
- 2026-05-20 17:40:55
@article{a6b3488b-14d5-48f3-b503-7b3782eb8708,
abstract = {{<p>Motivation Machine learning offers a powerful approach for building predictive models from high-dimensional molecular data. Omics technologies such as transcriptomics, proteomics, and metabolomics quantify thousands of molecules simultaneously, providing deep insights into disease biology. Integrating multiple modalities can enhance predictive performance, as shown in histology-omics and holter-omics applications. To support streamlined, reproducible, and user-friendly multimodal analytics, we developed Omics BioAnalytics, an R Shiny platform for unified analysis, integration, and interpretation of diverse omics datasets. Results Omics BioAnalytics performs late integration using ensembles of elastic net models trained independently on each modality, with predictions averaged across datasets. The platform provides interactive dashboards for metadata exploration, exploratory analyses, differential expression, gene set analysis, and biomarker discovery. Results are visualized through dynamic plots and downloadable reports, ensuring transparent and reproducible workflows. A unique feature is the integrated multimodal Alexa Skill, which enables voice-based querying and rapid visualization. Together, these web and voice-enabled tools offer accessible and reproducible multimodal analytics for biomedical researchers, supporting the discovery of molecular signatures, predictive biomarkers, and therapeutic targets. Availability and implementation All source code, public datasets, video walkthroughs, and the deployed application are available at: https://github.com/CompBio-Lab/omicsBioAnalytics.</p>}},
author = {{Dyce, Josh and Rieskamp, Lea and Tebbutt, Scott J. and McManus, Bruce M. and Singh, Amrit}},
issn = {{2635-0041}},
language = {{eng}},
number = {{1}},
publisher = {{Oxford University Press}},
series = {{Bioinformatics Advances}},
title = {{Omics BioAnalytics : an RShiny application for multimodal biomarker panel discovery and assessment}},
url = {{http://dx.doi.org/10.1093/bioadv/vbaf307}},
doi = {{10.1093/bioadv/vbaf307}},
volume = {{6}},
year = {{2026}},
}