Inferring Loss-of-Heterozygosity from Tumors Using SNP Arrays
Author Information
Author(s): Beroukhim Rameen, Lin Ming, Park Yuhyun, Hao Ke, Zhao Xiaojun, Garraway Levi A, Fox Edward A, Hochberg Ephraim P, Mellinghoff Ingo K, Hofer Matthias D, Descazeaud Aurelien, Rubin Mark A, Meyerson Matthew, Wong Wing Hung, Sellers William R
Primary Institution: Dana-Farber Cancer Institute
Hypothesis
Can a hidden Markov model accurately infer loss-of-heterozygosity (LOH) from unpaired tumor samples using high-density SNP arrays?
Conclusion
The developed method accurately identifies regions of LOH in tumors without the need for paired normal samples.
Supporting Evidence
- The method identified 99% of SNP markers correctly as either retention or loss.
- It accurately identified 81% of LOH regions, including 98% of regions greater than 3 megabases.
- The integration of copy number analysis helped distinguish LOH from allelic imbalance.
Takeaway
Scientists created a new way to find missing parts of genes in tumors without needing normal samples, which helps understand cancer better.
Methodology
A hidden Markov model was developed to analyze SNP data from unpaired tumor samples, considering SNP intermarker distances and heterozygosity rates.
Potential Biases
Potential bias due to the use of reference samples from different ethnicities affecting specificity.
Limitations
The method may produce false LOH calls in regions of linkage disequilibrium.
Participant Demographics
The study involved 34 prostate cancer samples.
Statistical Information
P-Value
p<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