Philippine Science Letters
vol. 6 | no. 1 | 2013
published online February 21, 2013


Effects of migration patterns on estimates of dispersal rates inferred from genotypic data: a simulationbased study focused on the software STRUCTURE

by Maia H. Malonzo1,2, Rachel G. Ravago-Gotanco1, Arturo O. Lluisma1,*

1Marine Science Institute, University of the Philippines, 1101 Quezon City
2Current Address: Department of Information and Computer Science, Aalto University School  of Science, Helsinki, Finland



Recent developments in the analysis of genetic data now make possible the direct measurement of migration rates through individual-based assignment methods at ecological time frames relevant for resource management. While several software implementing these assignment methods have been examined for accuracy under various conditions of spatial patterns and rates of gene flow and population size, previous analyses have not examined the effects of temporal variations in dispersal rate on assignment accuracy. In this study, we evaluated the assignment accuracy of the widely used software, STRUCTURE, using simulated genetic datasets generated to reflect two patterns of temporal variation in dispersal rate: a normal distribution and a negative binomial distribution, the latter reflecting a pattern of migration commonly observed in natural populations in which the movement of a large number of migrants into the sink population is a rare event. We also evaluated the accuracy of different assignment models and varying sample sizes. The results of the simulations suggest that at the mean migration rate of 5 individuals per generation, STRUCTURE exhibits greater assignment accuracy from a negative binomial distribution relative to a normal distribution at smaller sample sizes of 20-50 individuals. This however is attributed to greater population structure among populations in which migration followed a negative binomial distribution, and its effect on recovering more accurate assignments. At sample sizes of 100 to 200 individuals, assignment accuracy was similar for the two distributions. Increasing the sample size generally resulted in reduced specificity in classification. At the larger sample sizes, increasing numbers of false positives were recovered for both normal and negative binomial distribution patterns, likely due to the proportionally increased probability of sampling individuals with recent migrant ancestry. Incorporating prior population information into models of migrant inference resulted in higher levels of accuracy in detecting actual migrants (true positive assignments), and reducing false positive assignments.

* Corresponding author
Email Address:
Received: September 2, 2012
Revised: December 27, 2012
Accepted: December 27, 2012
Published: February 21, 2013
Editor-in-charge: Eduardo A. Padlan
Reviewer: Jingky P. Lozano-Kühne