A procedure for identifying homologous alternative splicing events

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Alternative splicing is crucial for gene function diversity.
• Experimental methods for studying alternative splicing are challenging, especially at high throughput.
• Bioinformatics tools are needed to predict functional roles of alternative splice isoforms.

Purpose of the Study:

• To develop a bioinformatics method for identifying homologous alternative splicing events.
• To enable functional annotation of alternative splice isoforms using conserved events.

Main Methods:

• A four-step procedure combining sequence searches and neural networks.
• Involves BLAST searches for isoform homologues, candidate event construction, neural network scoring, and result filtering.

Main Results:

• The method demonstrated high performance on manually annotated homologous events.
• Achieved an accuracy of 0.99, precision of 0.98, and sensitivity of 0.93.
• Specificity ranged from 0.81 to 0.91 when no candidates were found.

Conclusions:

• The developed method successfully identifies homologous alternative splicing events.
• This approach can facilitate the functional annotation of alternative splice isoforms.