Complete alternative splicing events are bubbles in splicing graphs.

Area of Science:

• Molecular Biology
• Genomics
• Bioinformatics

Background:

• Eukaryotic gene expression involves alternative splicing (AS), generating diverse transcript forms.
• Current methods often use pairwise comparisons, underestimating the complexity of AS events.
• High-throughput sequencing yields vast amounts of transcript data, including noisy expressed sequence tags (ESTs).

Purpose of the Study:

• To develop a novel computational method for predicting complex alternative splicing (AS) events.
• To accurately characterize the full diversity of AS events, including those with more than two alternatives.
• To analyze large-scale transcriptomic datasets, integrating fragmented and full-length sequence data.

Main Methods:

• A novel computational method was developed to predict AS events from transcript annotations.
• The method efficiently handles large datasets and incorporates noise reduction techniques.
• It integrates fragmented expressed sequence tag (EST) data with full-length complementary DNAs (cDNAs).

Main Results:

• Thousands of novel alternative splicing events were identified in human and murine transcript annotations.
• The majority of previously identified 'mutually exclusive' exons were found to participate in more complex splicing patterns.
• Analysis of complete AS events revealed a higher degree of splicing complexity than pairwise comparisons suggested.

Conclusions:

• The developed method enables dynamic updating of AS databases with new splice forms.
• Systematic analysis of complete AS events provides deeper insights into splicing mechanisms.
• The true extent of alternative splicing complexity is significantly underestimated by current analytical approaches.