Assessing Microarray Quality
Author Information
Author(s): Matthew N. McCall, Peter N. Murakami, Margus Lukk, Wolfgang Huber, Rafael A. Irizarry
Primary Institution: University of Rochester Medical Center
Hypothesis
How can we effectively assess the quality of Affymetrix GeneChip microarrays?
Conclusion
Approximately 10 percent of publicly available Affymetrix microarrays are of poor quality, and quality varies significantly across studies and labs.
Supporting Evidence
- The study found that single-array quality metrics are essential for clinical applications.
- The new GNUSE metric was shown to perform as well as existing multi-array metrics.
- Quality varied significantly across different studies and labs, indicating a need for better standards.
Takeaway
This study looked at how to check if microarray data is good or bad, and found that about 10 out of every 100 samples are not good enough to use.
Methodology
The study defined microarray quality, reviewed existing metrics, and proposed a new single-array quality metric called GNUSE, assessing over 22,000 arrays from the Gene Expression Omnibus.
Potential Biases
Some labs are more likely to produce poor quality arrays than others, which could bias results.
Limitations
The study primarily focused on Affymetrix microarrays and may not generalize to other platforms.
Participant Demographics
The study assessed data from 809 studies across various labs and tissue types.
Statistical Information
P-Value
<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website