Related Experiment Videos
Interlaboratory comparability study of cancer gene expression analysis using oligonucleotide microarrays
Kevin K Dobbin1, David G Beer, Matthew Meyerson
1Cancer Diagnosis Program, National Cancer Institute/NIH, Bethesda, MD 20894, USA. dobbinke@mail.nih.gov
Summary
Multi-laboratory microarray analysis is feasible for gene expression studies. Standardized protocols ensure comparable and reproducible data, supporting clinical translation of genomic findings.
Area of Science:
- Genomics
- Molecular Biology
- Cancer Research
Background:
- Clinical application of gene expression data requires large-scale confirmatory studies.
- Reproducibility and comparability of microarray data across laboratories are essential for reliable results.
Purpose of the Study:
- To assess the comparability and reproducibility of microarray data generated from multiple laboratories.
- To evaluate the feasibility of combining data from different sites for a large-scale lung adenocarcinoma profiling project.
Main Methods:
- Four laboratories analyzed blinded samples (tumor tissues, cell lines, purified RNA) using a common protocol.
- Affymetrix Human Genome U133A arrays were used for microarray analysis.
- Data comparability was assessed through correlation analyses and hierarchical clustering.
Main Results:
- High within- and between-laboratory correlations were observed across all sample types.
- Intraclass correlation was only slightly higher within than between laboratories.
- Hierarchical clustering demonstrated that samples clustered by origin, not by laboratory, indicating strong reproducibility.
Conclusions:
- Standardized protocols enable multi-laboratory microarray analysis with high data comparability.
- This approach is feasible for large-scale projects, facilitating the clinical translation of gene expression findings.
- Reproducibility is robust, even for genes with low expression levels and small variations.