Related Experiment Video
Updated: Jun 27, 2025
07:35
Selecting Multiple Biomarker Subsets with Similarly Effective Binary Classification Performances
Published on: October 11, 2018
7.5K
PerSEveML: a web-based tool to identify persistent biomarker structure for rare events using an integrative machine
Sreejata Dutta1, Dinesh Pal Mudaranthakam1,2, Yanming Li1,2
1Department of Biostatistics & Data Science, University of Kansas Medical Center, Kansas City, Kansas, USA. msardiu@kumc.edu.
Molecular Omics
|May 1, 2024
Summary
PerSEveML is a new web tool that uses machine learning to predict rare events in complex omics data. It helps researchers identify key biological features and generate new hypotheses from large datasets.
Area of Science:
- Bioinformatics
- Computational Biology
- Machine Learning
Background:
- Omics data analysis presents computational challenges due to high dimensionality and rare events.
- Existing machine learning (ML) methods are useful, but bioinformatics tools integrating ML for biological insight are limited.
Purpose of the Study:
- Introduce PerSEveML, an interactive web tool for predicting rare events in omics data.
- Enable feature selection and hypothesis generation by uncovering underlying biological structures.
Main Methods:
- Developed an integrative machine learning approach with crowd-sourced intelligence.
- Implemented evaluation metrics, entropy, and rank scores for feature analysis.
- Utilized a web-based interactive platform for user accessibility.
Main Results:
- PerSEveML successfully predicted rare events across diverse, biologically complex datasets.
- The tool generated valid hypotheses by organizing features into selected, unselected, and fluctuating categories.
- Demonstrated the utility of PerSEveML in understanding ML method contributions.
Conclusions:
- PerSEveML offers a novel bioinformatics solution for analyzing high-dimensional omics data with rare events.
- The tool facilitates hypothesis generation and biological discovery from complex datasets.
- PerSEveML enhances the interpretability of machine learning applications in biology.
