Related Experiment Video
Updated: Aug 1, 2025
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
6.9K
MutBLESS: A tool to identify disease-prone sites in cancer using deep learning
Medha Pandey1, M Michael Gromiha1
1Department of Biotechnology, Bhupat and Jyoti Mehta School of Biosciences, Indian Institute of Technology Madras, Chennai 600036, India.
Biochimica Et Biophysica Acta. Molecular Basis of Disease
|April 27, 2023
Summary
Identifying cancer-driving mutations is crucial but challenging. This study introduces a deep learning method to accurately predict disease-prone mutation sites, aiding in cancer research and therapeutic strategy development.
Area of Science:
- Genomics
- Bioinformatics
- Computational Biology
Background:
- Identifying driver mutations in cancer is experimentally costly and time-consuming.
- Understanding mutation impact across cancer stages is a key challenge in cancer biology.
Purpose of the Study:
- To develop an accurate computational method for predicting disease-prone mutation sites.
- To identify sequence-based features associated with cancer-driving mutations.
- To create a web server for public use in cancer research.
Main Methods:
- Collected and analyzed experimentally validated driver mutations from 22 cancer types.
- Utilized deep neural networks incorporating amino acid sequence features (physicochemical properties, secondary structure, tri-peptide motifs, conservation scores).
- Developed a web server (MutBLESS) for predicting cancer-prone sites.
Main Results:
- Identified distinct amino acid motifs (AAA, LR) prevalent in disease-prone sites and others (QPP, QF) in neutral sites.
- Achieved high prediction performance with an average AUC of 0.97 for specific cancer types (BRCA, LAML, EC, STAD, SKCM).
- Demonstrated excellent overall performance with average sensitivity (96.56%), specificity (97.39%), and accuracy (97.64%) for cancer-specific mutations.
Conclusions:
- The developed deep learning method effectively predicts disease-prone mutation sites.
- This approach can significantly assist in identifying cancer-specific mutations and developing targeted therapeutic strategies.
- The publicly available web server provides a valuable tool for the cancer research community.
