Detection of pre-mRNA involved in abnormal splicing using Graph Neural Network and Nearest Correlation Method.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• DNA contains genetic information in exons, separated by introns.
• Metabolic activities and environmental factors can damage DNA, leading to Single Nucleotide Variants (SNVs).
• Pathogenic SNVs can cause abnormal RNA splicing, resulting in genetic diseases like cancer.

Purpose of the Study:

• To develop a new method for detecting pre-mRNA with SNVs that lead to aberrant RNA splicing.
• To improve the accuracy of genetic disease diagnosis and treatment selection.

Main Methods:

• Utilized a Graph Neural Network (GNN) and the Nearest Correlation (NC) method.
• Extracted sequence features and calculated similarities using the NC method.
• Trained a GNN model on sequence similarities to classify pre-mRNA with SNVs, feeding into LightGBM.

Main Results:

• The method achieved a specificity of 0.73±0.080 and recall of 0.69±0.13.
• Evaluated performance using G-mean (0.71±0.073), ROC-AUC (0.76±0.062), and PR-AUC (0.42±0.10).

Conclusions:

• The proposed method aids in preventing misdiagnosis of diseases.
• It is expected to facilitate the selection of appropriate treatments for genetic diseases.