Significance Thresholds for Genomewide Association Studies
Author Information
Author(s): Frank Dudbridge, Arief Gusnanto
Primary Institution: MRC Biostatistics Unit, Institute for Public Health
Hypothesis
What significance threshold is appropriate for genomewide association studies?
Conclusion
The study estimates a genomewide significance threshold of about 7.2 × 10−8 for the UK Caucasian population.
Supporting Evidence
- The study used a large sample size of 1,485 blood donors and 1,504 members of the 1958 British Birth Cohort.
- Permutation testing was employed to estimate the significance threshold, preserving the correlation structure in the sample.
- The estimated genomewide significance threshold aligns with previous estimates from the International HapMap Consortium.
Takeaway
This study helps scientists understand how to set the right significance levels when looking for genetic links to diseases.
Methodology
The study used genotypes from the Wellcome Trust Case-Control Consortium and employed permutation testing to estimate significance thresholds.
Potential Biases
The study may be influenced by the prior odds of SNP associations being exaggerated.
Limitations
The results apply only to the UK Caucasian population and may not generalize to other populations.
Participant Demographics
The study included UK Caucasian participants, specifically 1,485 blood donors and 1,504 members of the 1958 British Birth Cohort.
Statistical Information
P-Value
7.2 × 10−8
Confidence Interval
(6.3–8.9) × 10−8
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website