Using ESTs to improve the accuracy of de novo gene prediction
2006
Improving Gene Prediction with ESTs
publication
Evidence: high
Author Information
Author(s): Wei Chaochun, Brent Michael R
Primary Institution: Washington University
Hypothesis
Can combining EST alignments with de novo gene prediction systems improve the accuracy of gene structure predictions?
Conclusion
The new systems TWINSCAN_EST and N-SCAN_EST are more accurate than their predecessors while still able to discover novel genes.
Supporting Evidence
- TWINSCAN_EST shows a 14% improvement in sensitivity and 13% in specificity for C. elegans.
- N-SCAN_EST produced a 6% improvement in sensitivity and 1% in specificity for the human genome.
- The new systems retain the ability to discover novel genes without EST alignments.
Takeaway
This study shows that using information from ESTs can help predict genes better, even those that haven't been sequenced yet.
Methodology
The study developed TWINSCAN_EST and N-SCAN_EST by integrating EST alignments with existing gene prediction systems.
Limitations
The accuracy improvements may not apply to all genes, especially those with no EST alignments.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website