Long runs of homozygosity (ROH) are often caused by "close" genetic relatedness between parents. ROH can be used to investigate inbreeding in populations, causes of complicated pedigree reconstruction and overall population sizes. ROH can also be viewed through two characteristics: (a) what proportion of the population yields no evidence of ROH, and (b) what is the mean ROH in a population?

So far, no hypothesis test exists for testing whether the above characteristics are either consistent with some "expected" parameters (one-sample tests), or whether two populations differ in these characteristics (two-sample tests).

Here I derive the required the distributions of such test statistics and suggest hypothesis tests to answer the above questions.

(Apologies, currently the code and method are private as the paper using this method is in review).