Related Experiment Video
Updated: May 24, 2026
07:15
Determining the Likelihood of Variant Pathogenicity Using Amino Acid-level Signal-to-Noise Analysis of Genetic Variation
Published on: January 16, 2019
Deep Learning-Based Prediction of Pathogenicity for ABL1 Protein Variants Using Sequence Representation.
Gülbahar Merve Şilbir1, Burçin Kurt2
1Department of Electricity and Automation/Robotics and Artificial Intelligence, Çarşıbaşı Vocational School, Trabzon University, Trabzon, Turkey.
Studies in Health Technology and Informatics
|May 23, 2026
Summary
This study uses deep learning to predict the pathogenicity of ABL1 single amino acid variants (SAVs). A Convolutional Neural Network model achieved high accuracy, aiding in understanding genetic disorders.
Area of Science:
- Genomics
- Computational Biology
- Bioinformatics
Background:
- The ABL1 gene encodes a non-receptor tyrosine kinase involved in leukemia and genetic disorders.
- Accurate prediction of single amino acid variant (SAV) pathogenicity is crucial for genetic diagnostics.
Purpose of the Study:
- To develop and evaluate a deep learning approach for predicting the pathogenicity of ABL1 SAVs using amino acid sequence representations.
- To compare the performance of various deep learning architectures for variant effect prediction.
Main Methods:
- Collected variant data from UniProt, ClinVar, dbSNP, and EVS databases.
- Encoded variant data using embedding and transformer-based methods.
- Evaluated Feedforward, Conv1D, BiLSTM, Transformer, and Prot-BERT architectures.
Main Results:
- The Convolutional Neural Network (CNN) model demonstrated superior performance.
- CNN achieved an AUC of 0.86, 93% accuracy, and 0.93 specificity in distinguishing benign from pathogenic variants.
- The study identified sequence-centered deep learning as a powerful approach for variant effect prediction.
Conclusions:
- Deep learning frameworks, particularly CNNs, show significant potential for accurate variant pathogenicity prediction.
- AI-assisted tools can enhance computational genomics and genetic disorder analysis.
- This work supports the integration of machine learning in understanding the functional impact of genetic variants.
