Predicting Microarray Cross-Hybridization
Author Information
Author(s): Chen Yian, Chou Cheng-Chung, Lu Xinghua, Slate Elizabeth H, Peck Konan, Xu Wenying, Voit Eberhard O, Almeida Jonas S
Primary Institution: Medical University of South Carolina
Hypothesis
Can a multivariate model improve the prediction of cross-hybridization in microarrays?
Conclusion
A systematic multivariate approach can effectively assess the importance of multiple sequence features for hybridization.
Supporting Evidence
- The most contiguous base pairs were found to be the best predictor of hybridization.
- Inclusion of target GC content improved predictive power.
- Multivariate models outperformed univariate models in predicting hybridization.
Takeaway
This study helps scientists understand how different DNA sequences stick together in experiments, which can help avoid mistakes in reading results.
Methodology
The study used multiple linear regression, regression trees, and artificial neural networks to analyze hybridization data from DNA microarrays.
Potential Biases
Potential biases may arise from the small sample size and the specific conditions under which the experiments were conducted.
Limitations
The model development was based on a small dataset, which may limit the generalizability of the findings.
Statistical Information
P-Value
<10-4
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website