Determinação do tamanho mínimo amostral para estudos genéticos populacionais em plantas, utilizando marcadores microssatélites

Abstract

When starting a population genetic study, the researcher must first decide the number of individuals to be used for each population. For animals, this question seems to be clear. However, for plants, studies with real data to determine the influence of sample size on genetic indices are limited, being almost limited to Gymnosperm. To help answer this question, four species of Angiosperm were evaluated in this study: a large tree (Campomanesia xanthocarpa O. Berg), a medium tree (Cecropia pachystachya Trécul), a small tree (Psidium guajava L.) and an subshrub (Baccharis crispa Spreng.). From a set of real data of 100 individuals of each species, obtained with microsatellite markers, 50 repetitions were obtained by random for the sample sizes of 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 75 and 95 individuals, totaling 2400 subpopulations (600 for each species). For each subpopulation, the number of alleles (Na), expected heterozygosity (HE), allelic frequency and coefficient of genetic differentiation (FST) were calculated. With the data obtained, the range of data, the median, and the quartiles were determined. The results showed that, for Na, the sample size of 45 was able to detect 90% of alleles in C. xanthocarpa and B. crispa, and 35 individuals in C. pachystachya and P. guajava. For HE, the sample size of 30 individuals is sufficient to detect genetic estimates for three of the four species studied. The sample size of 40 individuals is sufficient for detection in all species of alleles that presented frequencies above 0.05 in the original population. On the other hand, rare alleles are not detectable in all subpopulations in any sample size. For the FST from the sample size of 10 individuals, the difference between the original population and the subpopulations was less than 0.05 in C. xanthocarpa and C. pachystachya; and 15 for P. guajava and B. crispa. In the analysis of rare alleles, no differences were found in genetic indices and structure between populations of 100 and 30 individuals with and without rare alleles. The overall analysis of the data shows that, from the sample size of 30 individuals, the data can already be considered reliable, as most indices already show values close to those observed in the original population. Studies with a number of individuals smaller than 30 should only be carried out when the population size does not allow reaching the ideal sample size.