Related Experiment Video
Updated: Jun 24, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
1.0K
Prediction of miRNAs and diseases association based on sparse autoencoder and MLP
Si-Lin Sun1, Bing-Wei Zhou1, Sheng-Zheng Liu1
1Department of Information Science Technology, Hainan Normal University, Haikou, Hainan, China.
Frontiers in Genetics
|June 14, 2024
Summary
We developed a novel deep learning method (SPALP) to predict associations between microRNAs (miRNAs) and diseases. This approach significantly improves prediction accuracy, aiding in understanding disease mechanisms and developing new treatments.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- MicroRNAs (miRNAs) are small, non-coding RNA molecules with critical regulatory functions.
- Aberrant miRNA expression is increasingly linked to various human diseases.
- Understanding miRNA-disease associations is vital for elucidating disease pathogenesis and identifying therapeutic targets.
Purpose of the Study:
- To propose and validate a novel computational method for predicting miRNA-disease associations.
- To leverage deep learning techniques for enhanced accuracy in identifying these relationships.
- To provide a tool for discovering potential biomarkers and therapeutic strategies related to miRNAs and diseases.
Main Methods:
- A sparse autoencoder and multi-layer perceptron (MLP) method (SPALP) was developed.
- Sparse autoencoder was used for feature learning and extracting latent features of miRNAs and diseases.
- Latent features were integrated with miRNA functional similarity and disease semantic similarity data to build comprehensive datasets for MLP-based prediction.
Main Results:
- The SPALP method achieved a high prediction accuracy of 94.61% and an AUC value of 0.9859.
- Comparative experiments demonstrated superior performance over traditional and existing deep learning methods.
- Case studies showed successful prediction of top miRNAs associated with five common elderly diseases, with high validation rates.
Conclusions:
- The SPALP approach effectively predicts miRNA-disease associations, offering a robust tool for bioinformatics research.
- This method addresses the challenges of analyzing large-scale omics data.
- SPALP contributes to a deeper understanding of miRNA roles in disease progression and potential clinical applications.
