Posting a question received by email from a user:

"I have several questions about the functions in PLINK which I haven't found in GWASTools.

1. Does GWASTools provide genomic control feature?

2. Also how about FDR?

3. In plink, there is a model called genotypic model, basically it is a 2 degree of freedom test with two parameters b_{snp=1} and b_{snp=2}. I wonder does GWASTools provide this model?"

1. Does GWASTools provide genomic control feature?

Genomic control is a method for dealing with population substructure in an association test. (See for example http://en.wikipedia.org/wiki/Population_stratification#Genomic_Control and http://www.ncbi.nlm.nih.gov/pubmed/11855950.) This feature is not available in GWASTools; instead, I recommend the now-common practice of including eigenvectors from Principal Component Analysis as covariates in the model. See the Data Cleaning vignette for an example of how to calculate these eigenvectors using the SNPRelate package and include them in a GWASTools analysis.

2. Also how about FDR?

FDR stands for False Discovery Rate. PLINK association output includes two columns of FDR-corrected p-values using the methods of Benjamini & Hochberg (1995) and Benjamini & Yekutieli (2001). You can apply these corrections in R using p.adjust with method="BH" or "BY". However, the more common practice is to leave the p-values as-is and apply a threshold of 5e-8 for genome-wide significance.

3. In plink, there is a model called genotypic model, basically it is a 2 degree of freedom test with two parameters b_{snp=1} and b_{snp=2}. I wonder does GWASTools provide this model?"

The conventional allelic test considers the effect of one allele (A or B) versus the other; while the genotypic test considers the effects of the three genotypes (AA, AB, BB) independently. Box 2 of this paper (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3154648/) gives a good summary. This model is not available in GWASTools.