Identifying a Common Cancer Signature from Microarray Data
Author Information
Author(s): Xu Lei, Geman Donald, Winslow Raimond L
Primary Institution: Johns Hopkins University
Hypothesis
Can large-scale integration of cancer microarray data identify a robust common cancer signature?
Conclusion
A robust common cancer signature consisting of 46 genes has been identified, which may be useful for cancer diagnosis.
Supporting Evidence
- The TSPG classifier achieved high accuracy (> 85%) on 19 of 26 data sets.
- The common cancer signature was validated on independent test data sets.
- Statistical significance was observed in classification accuracy across multiple cancer types.
Takeaway
Researchers combined data from many cancer studies to find a set of genes that can help tell if someone has cancer.
Methodology
The study used the Top-Scoring Pair of Groups (TSPG) classifier and repeated random sampling to analyze integrated microarray data.
Potential Biases
Potential bias due to the reliance on specific microarray platforms.
Limitations
The method can only integrate microarray data from the same platforms, which may limit the analysis.
Participant Demographics
Data included samples from 21 major human cancer types.
Statistical Information
P-Value
p<0.03
Statistical Significance
p<0.03
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website