The impact of microsatellite electromorph size homoplasy on multilocus population structure estimates in a tropical tree (Corythophora alta) and an anadromous fish (Morone saxatilis)

Authors

  • RACHEL I. ADAMS,

    1. Georgetown University, Department of Biology, Reiss Sciences Building 406, 37th and O Streets NW, Washington, DC 20057–1229 USA;
    Search for more papers by this author
  • KYLE M. BROWN,

    1. Georgetown University, Department of Biology, Reiss Sciences Building 406, 37th and O Streets NW, Washington, DC 20057–1229 USA;
    Search for more papers by this author
  • MATTHEW B. HAMILTON

    Corresponding author
    1. Georgetown University, Department of Biology, Reiss Sciences Building 406, 37th and O Streets NW, Washington, DC 20057–1229 USA;
    2. Biological Dynamics of Forest Fragments Project, National Institute for Research in the Amazon, Manaus, Brazil
    Search for more papers by this author

Matthew B. Hamilton. Fax: 1 202-687-5662; E-mail: hamiltm1@georgetown.edu

Abstract

Microsatellite allelic states are determined by electrophoretic sizing of polymerase chain reaction fragments to define electromorphs. Numerous studies have documented that identical microsatellite electromorphs are potentially heterogeneous at the DNA sequence level, a phenomenon called electromorph size homoplasy. Few studies have examined the impact of electromorph size homoplasy on estimates of population genetic parameters. We investigated the frequency of microsatellite electromorph size homoplasy for 12 loci in the tropical tree Corythophora alta and 11 loci in the anadromous fish Morone saxatilis by sequencing 14–23 homozygotes per locus sampled from multiple populations for a total of 453 sequences. Sequencing revealed no homoplasy for M. saxatilis loci. Seven C. alta loci exhibited homoplasy, including single and compound repeat motifs both with and without interruptions. Between 12.5 and 42.9% of electromorphs sampled per locus showed size homoplasy. Two methods of correction for homoplasy in C. alta generally produced little or no change in single-locus estimates of RST, except for two loci in which some additional differentiation among populations was revealed. Twelve-locus estimates of RST (including the seven loci corrected for homoplasy) were slightly greater than estimates from uncorrected data, although the 95% confidence intervals overlapped. The frequency of methodological errors such as clerical mistakes or sample mislabelling per genotype scored was estimated at 5.4 and 7.3% for C. alta and M. saxatilis, respectively. Simulations showed that the increase in RST produced by homoplasy correction was only slightly larger than variation in RST estimates expected to be caused by methodological errors.

Ancillary