Identifying Homologous Alternative Splicing Events
Author Information
Author(s): David Talavera, Modesto Orozco, Xavier de la Cruz
Primary Institution: Institut de Recerca Biomèdica (IRB), Parc CientÃfic de Barcelona
Hypothesis
Can a method be developed to accurately identify homologous alternative splicing events using neural networks and sequence searches?
Conclusion
The method allows for the identification of homologous alternative splicing events with a high success rate, indicating its potential for functional annotation.
Supporting Evidence
- The method showed an accuracy of 0.99 when tested on 473 manually annotated pairs of homologous events.
- Precision was reported at 0.98 and sensitivity at 0.93.
- The method outperformed a control method that did not use neural networks.
Takeaway
The researchers created a way to find similar gene variations that happen when genes are spliced differently, which can help us understand how genes work.
Methodology
The method involves a four-step process including BLAST searches, candidate event construction, scoring with neural networks, and filtering results.
Potential Biases
Potential bias may arise from the similarity between isoforms and the nature of the alternative splicing events.
Limitations
The method may not identify homologous events if no candidates are available in the isoform database.
Statistical Information
P-Value
0.01
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website