In studying the effect of a SNP on susceptibility to a disease in a case-control design, How I can calculate the detectable OR for a given sample size with a power of 80%? which parameters I need and how can I find them. which software do u recommend?
The sample size with sufficient statistical power is critical to accomplish the genetic risk variant association in human complex diseases such as CAD, Diabetes, Hypertension, Cancer, etc. Analysis expending a large sum of SNPs requires a higher sample size, theoretically testing a single SNP requires 248 cases, while screening 500K – 1 million SNPs requires 1,206 & 1,255 cases respectively. These case numbers were assumed based on the following criteria:
I. Odds ratio of 2,
II. 5% disease prevalence,
III. 5% minor allele frequency,
IV. Complete linkage disequilibrium (LD),
V. 1:1 case/control ratio, and
VI. 5% error rate in an allelic test.
The statistical power of 80% is widely used to avoid false negative associations and to determine a cost-effective sample size in large-scale association studies.
Number of cases required to achieve 80% power according to the different genetic models in a case-control study (Adapted from Hong and Park doi:10.5808/gi.2012.10.2.117)
Genetic Model ORhet/ORhomo
1.3/1.6 1.5/2.0 2.0/3.0 2.5/4
Allelic 1,974 789 248 134
Dominant 606 258 90 53
Co-dominant 2,418 964 301 161
Recessive 20,294 8,390 2,776 1,536
The Following links may be useful for you
http://osse.bii.a-star.edu.sg/
http://biostats.usc.edu/Quanto.html
All the best
Dear Davide
Thank you, but there is no power calculation function in SNPasscoc package.
Dear Javad,
I suppose you would like PGA (http://dceg.cancer.gov/tools/analysis/pga) and Quanto software (http://biostats.usc.edu/software).
Cheers
The sample size with sufficient statistical power is critical to accomplish the genetic risk variant association in human complex diseases such as CAD, Diabetes, Hypertension, Cancer, etc. Analysis expending a large sum of SNPs requires a higher sample size, theoretically testing a single SNP requires 248 cases, while screening 500K – 1 million SNPs requires 1,206 & 1,255 cases respectively. These case numbers were assumed based on the following criteria:
I. Odds ratio of 2,
II. 5% disease prevalence,
III. 5% minor allele frequency,
IV. Complete linkage disequilibrium (LD),
V. 1:1 case/control ratio, and
VI. 5% error rate in an allelic test.
The statistical power of 80% is widely used to avoid false negative associations and to determine a cost-effective sample size in large-scale association studies.
Number of cases required to achieve 80% power according to the different genetic models in a case-control study (Adapted from Hong and Park doi:10.5808/gi.2012.10.2.117)
Genetic Model ORhet/ORhomo
1.3/1.6 1.5/2.0 2.0/3.0 2.5/4
Allelic 1,974 789 248 134
Dominant 606 258 90 53
Co-dominant 2,418 964 301 161
Recessive 20,294 8,390 2,776 1,536
The Following links may be useful for you
http://osse.bii.a-star.edu.sg/
http://biostats.usc.edu/Quanto.html
All the best
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis