Table 4.

Major assumptions of and questions relating to descriptive statistics and their implications in population genetic studies using microsatellites

Method	Assumption or question	References	Related issues	References
In general	Diploid genome	Dufresne et al. (2014)	For haploids or especially polyploids, options (in terms of both statistics and packages) are fewer	Dufresne et al. (2014)
F-statistics	Biallelic markers1 under infinite allele model of mutation. Often estimated using heterozygosity	Holsinger and Weir (2009), Meirmans and Hedrick (2011)	FST and relatives depressed under high diversity, requiring adjusted versions, especially problematic when mutation rate exceeds migration rate	Hedrick (2005), Jakobsson et al. (2013)
			Requires use of unbiased estimators of heterozygosity. Cannot compare subpopulations or loci with different levels of gene diversity	Meirmans and Hedrick (2011)
F-statistics	Infinite island model of population structure	Meirmans and Hedrick (2011)	Violation of migration assumptions require FST to be estimated pairwise or using alternatives (extension of θ, or F-model)	Weir and Hill (2002), Gaggiotti and Foll (2010)
			Correlation of allele frequencies among populations can cause overestimation using most methods	Fu et al. (2005)
RST, ρST fire	Stepwise mutation model (SMM)	Chakraborty and Nei (1982), Slatkin (1995), Rousset (1996)	Likely confounded by deviations from SMM. Levels of diversity and structure in the sample likely influence relative performance of FST and RST	Balloux et al. (2000), Balloux and Goudet (2002), Balloux and Lugon-Moulin (2002)
			Parameters including allele size are associated with high variance; should be estimated using analysis of molecular variance (AMOVA)	Michalakis and Excoffier (1996), Balloux and Goudet (2002)
D	Depends only on allelic differentiation	Jost (2008)	Can be sensitive to markers with high mutation rates	Leng and Zhang (2011, 2013)
In general	Which statistics should be employed?	–	Recommended to report as many as possible with microsatellites and ensure clarity in distinguishing parameters and estimators, such as for RST, ρST, or ΦST	Heller and Siegismund (2009)
In general	How should F-statistics be interpreted?	–	For any parameter: Microsatellites may substantially underestimate population structure, and interpretation has been described as “dangerous.”	Balloux and Lugon-Moulin (2002), Leng and Zhang (2011)

G_ST and some implementations of θ can be used for multiallelic markers.