Predicting the change of exon splicing caused by genetic variant using support vector regression

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Genetic variants are linked to diseases like cancer.
• Understanding how genetic variations affect alternative splicing is crucial for disease pathogenesis research.

Purpose of the Study:

• To develop a novel computational approach, PredPSI-SVR, for predicting the impact of genetic variants on alternative splicing, specifically exon skipping events.
• To enhance the understanding of variant-induced disease mechanisms.

Main Methods:

• Developed PredPSI-SVR using support vector regression.
• Extracted 42 features from exon sequences and flanking regions.
• Employed a greedy feature selection algorithm to identify the eight most predictive features.

Main Results:

• Achieved a Pearson correlation coefficient (PCC) of 0.570 in 10-fold cross-validation on the vex-seq training dataset.
• Ranked 2nd in the vex-seq blind test with a PCC of 0.566, demonstrating robustness.
• Showed utility in prioritizing deleterious synonymous mutations.

Conclusions:

• PredPSI-SVR is a robust and effective method for predicting the impact of genetic variants on alternative splicing.
• The approach aids in understanding disease pathogenesis and prioritizing genetic variants of concern.
• The method is publicly available for research use.