hierfstat, a package for r to compute and test hierarchical strongfstrong.pdfVIP

Hierfstat, a package for R to compute and test hierarchical F-statistics J´erˆome Goudet Dept. of Ecology and Evolution BB, UNIL, CH-1015 Lausanne, Switzerland jerome.goudet@unil.ch 16th August 2004 Abstract The package Hierfstat for the statistical software R (R Development Core Team [2003]) allows to estimate hierarchical F-statistics from a hierar- chy with any numbers of levels. In addition, it allows testing the statistical significance of population differentiation for these different levels, using a generalized likelihood-ratio test. The package Hierfstat is available from http://www.unil.ch/popgen/softwares/hierfstat.htm. Population biologists have a long standing interest in estimating population structure with the help of genetic markers. To this end, F-statistics (Wright [1951], Nei [1987],Weir [1996]) are very commonly used and many computer pro- grams are available to estimate these quantities (see for instance an extensive list at http://www.bio.ulaval.ca/louisbernatchez/links.htm#soft_pop_struct). F-statistics are usually defined for a two level hierarchy, where individuals (usu- ally diploids) are located in sub-populations. From such a hierarchy, the statistic FIS , where I stands for individual and S for sub-population, estimates the de- parture from panmixia at the level of the sub-populations, while FST (where T stands for total) quantifies the differences in allele frequencies among populations. This two level hierarchy has often been found to be too restrictive compared to the biology of the species under investigation. Patches are often nested in localities, themselves potentially nested in areas again nested in