Using Elastic Net in Canonical Correlation Analysis for Gene Expression and DNA Markers
Author Information
Author(s): Waaijenborg Sandra, Zwinderman Aeilko H
Primary Institution: Academic Medical Center
Hypothesis
Inter-individual variation in gene expression levels can be quantified through the association with DNA markers using multivariate techniques.
Conclusion
The adaptation of the elastic net to canonical correlation analysis simplifies the interpretation of components while maintaining information integrity.
Supporting Evidence
- The method reduces the number of variables in CCA components without significant loss of information.
- Strong intraset correlation is accounted for by the grouping effect of the ridge penalty.
- The penalized CCA is effective in situations where the number of variables exceeds the number of subjects.
Takeaway
This study shows a way to understand how genes and DNA markers are related by using a special math method that makes it easier to see patterns.
Methodology
The study used penalized canonical correlation analysis with the elastic net to analyze gene expression and DNA marker data from 194 subjects.
Limitations
The study focused only on the first set of CCA components and did not explore the second set despite its high canonical correlation.
Participant Demographics
Data from 14 three-generation CEPH families, including grandparents, parents, and children.
Want to read the original?
Access the complete publication on the publisher's website