Related Experiment Videos
Artificial neural network prediction of antisense oligodeoxynucleotide activity
Michael C Giddings1, Atul A Shah, Sue Freier
1Department of Human Genetics, University of Utah, SLC, UT 84112, USA and. Isis Pharmaceuticals, Carlsbad, CA 92008, USA. giddings@unc.edu
Nucleic Acids Research
|October 5, 2002
Summary
Identifying effective antisense oligodeoxynucleotide targets is challenging. A new artificial neural network system predicts effective targets with 53% accuracy, potentially reducing in vivo screening by fivefold.
Area of Science:
- Molecular Biology
- Bioinformatics
- Computational Biology
Background:
- Antisense oligodeoxynucleotides (oligos) target mRNA, but identifying effective sequences is difficult.
- Previous research linked short sequence motifs to oligo activity.
- A predictive model for oligo efficacy is needed.
Purpose of the Study:
- To develop a predictive artificial neural network system for mapping tetranucleotide motif content to antisense oligo activity.
- To improve the identification of efficacious antisense oligodeoxynucleotide target sites.
- To reduce the trial-and-error process in selecting active oligos.
Main Methods:
- Developed a predictive artificial neural network system based on tetranucleotide motif content.
- Trained the system for high-specificity prediction.
- Cross-validated the system against literature (348 oligos) and proprietary (908 oligos) databases.
Main Results:
- The system achieved 53% accuracy in identifying effective oligos (reducing mRNA expression to <25%).
- This contrasts with <10% success rates for traditional trial-and-error methods.
- A potential fivefold reduction in in vivo screening for active oligos was suggested.
Conclusions:
- The developed artificial neural network system effectively predicts antisense oligodeoxynucleotide efficacy.
- This approach significantly improves upon traditional methods for target site identification.
- A web interface is available for predicting effective oligo targets in RNA transcripts.