Department of Ecology & Evolution, Biology Building, UNIL, CH-1015 LAUSANNE, Switzerland
Tel :+41(0)21 692 4160 (sec) ; Fax:+41(0)21 692 4265



A software developed by Jérôme Goudet.

FSTAT is a computer package for PCs which estimates and tests gene diversities and differentiation statistics from codominant genetic markers. It computes both Nei and Weir & Cockerham families of estimators of gene diversities and F-statistics, and tests them using randomisation methods. The current version is 2.9.4. Previous versions are not maintained anymore.

To install FSTAT 2.9.4, extract file once you have downloaded it. This will create under the folder you have chosen the following architecture of folders and files:

The following folders contain data and example files, use of which is detailed in the help file:
.\data\bias dispersal
.\data\multiple regression

What FSTAT does

From a data set of codominant or haploid genetic markers, FSTAT calculates the following:

- Number of individuals per sample and loci.
- Allele frequency estimated per sample and overall.
- Observed and expected number of each genotype per sample and locus.
- Unbiased gene diversity per sample and locus.
- Number of alleles sampled per locus and sample, as well as overall.
- [NEW] Allelic richness per locus and sample, as well as overall samples
- Fis per locus and sample, as well as a test of whether it is significantly positive or negative (significant deficit and excess of heterozygotes respectively).
- Nei's (1987) estimators of gene diversities and differentiation.
- Weir & Cockerham (1984) Capf (Fit), theta (Fst) and smallf (Fis) estimated per allele, per locus and overall. FSTAT also calculates Hamilton's (1971) relatedness relat=2Fst/(1+Fit), calculated using an estimator strictly equivalent to Queller and Goodnight's (1989). This measure is the average relatedness of individuals within samples when compared to the whole data set.

Confidence intervals based on resampling schemes are provided for Weir & Cockerham statistics:

- Jackknifing per locus over samples is performed. The resampling unit are the different samples. This procedure is only carried out if there are more than 4 samples.
- Jackknifing over loci. In this case, the resampling units are the different loci. This procedure is only carried out if there are more than 4 loci.
- Bootstrapping over loci. Bootstrapping over loci is performed only when there is more than 4 loci in the data set.

- Estimation of R- statistics (Slatkin, 1995), specifically designed for microsatellite undergoing stepwise mutations.
- Estimation of Fst (theta) per pair of samples.

- Overall test whether each sample at each locus is in HW equilibrium.
- Test whether the entire data set is in HW equilibrium.
- Test whether samples are differentiated assuming either that there is HW within samples or that there is not (Only one of these 2 tests can be carried out).
- Test whether each sample at each locus is in Hardy-Weinberg (HW) equilibrium.
- Test whether each pair of samples is differentiated. The tests do not assume random mating within samples. A table of significant pairwise differentiation after corrections for multiple testing is produced.

- [NEW] Test whether each pair of loci in each sample and overall is at genotypic equilibrium

- [NEW] Test whether groups of samples differ for a large panel of statistics
- [NEW] Test whether categories of individuals differ in dispersal rates (four different tests)

- [NEW] Convert FSTAT format to GENEPOP and vice- versa
- [NEW] Performs multiple regression or Partial Mantel tests

Download FSTAT

Version 2.9.4

zipped archive file

Register as a user


Email address:



last modified on 23-08-2005 ; contact webmaster