Related Experiment Video
Updated: Jun 26, 2026
08:38
Targeted DNA Methylation Analysis by Next-generation Sequencing
Published on: February 24, 2015
Mining knowledge for the methylation status of CpG islands using alternating decision trees
Matthew B Carson1, Robert Langlois, Hui Lu
1Bioinformatics Program, Department of Bioengineering, University of Ilinois at Chicago, IL 60607, USA.
Summary
This study predicts CpG island methylation status on human chromosome 21 using sequence patterns. Machine learning models accurately identified methylated and unmethylated CpG islands, revealing key sequence characteristics.
Area of Science:
- Genomics
- Epigenetics
- Bioinformatics
Background:
- CpG island (CpGI) methylation is a crucial epigenetic modification in eukaryotes, regulating gene expression.
- Aberrant CpGI methylation is linked to various diseases and influences drug sensitivity.
- Understanding methylation patterns is vital for disease research and therapeutic development.
Purpose of the Study:
- To predict the methylation status of CpG islands (CpGIs) in human chromosome 21.
- To identify sequence patterns associated with methylated and unmethylated CpGIs.
- To gain insights into the factors influencing CpGI methylation.
Main Methods:
- Utilized sequence patterns to predict CpGI methylation status.
- Employed C4.5 algorithm with bagging and cost-sensitive learning for classification.
- Constructed 1000 alternating decision trees using bootstrapping to analyze conserved nodes.
Main Results:
- Achieved 85.6% accuracy, 82.8% sensitivity, and 86.4% specificity in predicting methylation status.
- Identified specific combinations of sequence patterns that differentiate between methylated and unmethylated CpGIs.
- Found conserved decision tree nodes indicative of methylation-associated sequence characteristics.
Conclusions:
- Sequence patterns can effectively predict CpGI methylation status on human chromosome 21.
- The study provides insights into the sequence-based determinants of CpGI methylation.
- This approach aids in understanding epigenetic regulation and its role in disease.
