Sex, size and timing: Sampling design for reliable population genetics analyses using microsatellite data
|Author(s)||Dubois Quentin1, Lebigre Christophe1, 2, Schtickzelle Nicolas1, Turlure Camille1|
|Affiliation(s)||1 : Catholic Univ Louvain, Earth & Life Inst, Biodivers Res Ctr, Louvain La Neuve, Belgium.
2 : IFREMER Ctr Bretagne, Unite Sci & Tech Halieut, Plouzane, France.
|Source||Methods In Ecology And Evolution (2041-210X) (Wiley), 2018-04 , Vol. 9 , N. 4 , P. 1036-1048|
|WOS© Times Cited||1|
|Keyword(s)||butterfly, conservation, genetic diversity, metapopulation, number of samples, period of sampling, population structure, sampling scheme optimization, sex ratio|
Population genetics is used in a wide variety of fields such as ecology and biodiversity conservation. How estimated genetic characteristics of natural populations can be influenced by the sampling design has been a long-standing concern. Multiple simulation and empirical studies illustrated the influence of both sample size and polymorphism of markers. However, our review of studies on butterfly population genetics indicates no consensus on sample size for the estimation of genetic diversity or differentiation. Furthermore, other aspects of sampling design (sex ratio and timing of sampling) were not addressed and their potential impact on genetic parameter estimates rarely explored.
Using a large empirical dataset (with spatial and temporal replicates) collected on a butterfly species, Boloria aquilonaris, as well as simulated datasets reflecting (1) three scenarios of migration–genetic drift equilibrium and (2) one scenario of parameter stabilization after 100,000 generations, we quantified the impacts of three aspects of genetic sampling design (namely, sample size, sex ratio and timing of sampling) on the estimation of allele frequencies and its potential downstream impact on the estimation of genetic parameters.
With empirical data, we found that sample size and timing of sampling strongly affected the accuracy of allele frequencies and the downstream analyses, while sex ratio did not. Our results were consistent across spatial and temporal replicates. Also, with simulated data, we showed that the genetic sampling design had limited effect in systems where dispersal outweighs genetic drift, while it can have major consequences on our understanding of the genetic diversity and population differentiation in systems dominated by genetic drift (such as most study systems with conservation concerns).
We advocate for careful consideration of all aspects of the sampling design in population genetics studies, that is a sufficient number of samples, while ensuring similar sex ratio among sampling locations and collecting with timing appropriate to the question under study. This is particularly important when the study aims at species conservation