Normalization of Microarray Expression Data and Its Impact on Linkage Analysis
Author Information
Author(s): Kim Yoonhee, Doan Betty Q, Duggal Priya, Bailey-Wilson Joan E
Primary Institution: National Human Genome Research Institute, National Institutes of Health
Hypothesis
Linkage analysis that uses expression trait data normalized by ignoring pedigree membership could improperly 'correct' the trait values.
Conclusion
Normalization within pedigrees generally increased familial correlations but decreased the strength of maximum LOD scores for linked traits.
Supporting Evidence
- 39% of traits had a p-value < 0.001 when comparing normalization methods.
- 95% of genes had a p-value < 0.001 when normalized using four individuals as their own pool versus independent individuals.
- Familial correlation coefficients were generally higher in data normalized within pedigrees.
Takeaway
This study looked at how to best adjust gene expression data from families to get accurate results. It found that normalizing data within families helps show true relationships but might make some results less strong.
Methodology
The study compared two normalization methods: RMA normalization using all individuals versus within-pedigree pools, analyzing gene expression data from 194 individuals across 14 CEPH Utah pedigrees.
Potential Biases
Normalization using all individuals may inflate linkage signals by assuming unrelated individuals share the same distribution.
Limitations
The study did not evaluate linkage of the 18 traits to all markers on other chromosomes, limiting the assessment of potential linkage signal inflation.
Participant Demographics
194 individuals from 14 CEPH Utah pedigrees.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.001
Want to read the original?
Access the complete publication on the publisher's website