On 2013 Jul 10, Matthew Stephens commented:

I'm interested in whether anyone has explored connections between the ideas here, and the "empirical null" ideas of Efron (eg Large-scale simultaneous hypothesis testing: the choice of a null hypothesis, JASA 2004) http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.179.2729

It seems the two have a similar goal - here one of the ideas (and probably the one that has most influence in practice, since it is now widely used) is to use the median of the test statistics to estimate an inflation factor (lambda). Efron also uses the test statistics near the center of the empirical distribution to do a similar thing, but in a different way. It might be interesting to compare the two approaches.

On 2013 Jun 18, Jill Barnholz-Sloan commented:

This paper proposes a computationally simple approach for adjustment for potential population stratification bias in case control studies. Genomic control and structured association methods, such as STRUCTURE, are widely used to test for and adjust for potential population stratification bias. Genomic control has been compared and contrasted with STRUCTURE (http://www.ncbi.nlm.nih.gov/pubmed/18852890)

On 2013 Jun 18, Jill Barnholz-Sloan commented:

This paper proposes a computationally simple approach for adjustment for potential population stratification bias in case control studies. Genomic control and structured association methods, such as STRUCTURE, are widely used to test for and adjust for potential population stratification bias. Genomic control has been compared and contrasted with STRUCTURE (http://www.ncbi.nlm.nih.gov/pubmed/18852890)

On 2013 Jul 10, Matthew Stephens commented:

I'm interested in whether anyone has explored connections between the ideas here, and the "empirical null" ideas of Efron (eg Large-scale simultaneous hypothesis testing: the choice of a null hypothesis, JASA 2004) http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.179.2729

It seems the two have a similar goal - here one of the ideas (and probably the one that has most influence in practice, since it is now widely used) is to use the median of the test statistics to estimate an inflation factor (lambda). Efron also uses the test statistics near the center of the empirical distribution to do a similar thing, but in a different way. It might be interesting to compare the two approaches.