Refereed Article Predicting the probability of outbreeding depression
Citation: Frankham R, Ballou JD, Eldridge MDB, Lacy RC, Ralls K, Dudash MR and Fenster CB . 2011. Predicting the probability of outbreeding depression. Conservation Biology. 25. (3): 465-475.Abstract:
Fragmentation of animal and plant populations typically leads to genetic erosion and increased probability of extirpation. Although these effects can usually be reversed by reestablishing gene flow between population fragments, managers sometimes fail to do so due to fears of outbreeding depression (OD). Rapid development of OD is due primarily to adaptive differentiation from selection or fixation of chromosomal variants. Fixed chromosomal variants can be detected empirically. We used an extended form of the breeders’ equation to predict the probability of OD due to adaptive differentiation between recently isolated population fragments as a function of intensity of selection, genetic diversity, effective population sizes, and generations of isolation. Empirical data indicated that populations in similar environments had not developed OD even after thousands of generations of isolation. To predict the probability of OD, we developed a decision tree that was based on the 4 variables from the breeders’ equation, taxonomic status, and gene flow within the last 500 years. The predicted probability of OD in crosses between 2 populations is elevated when the populations have at least one of the following characteristics: are distinct species, have fixed chromosomal differences, exchanged no genes in the last 500 years, or inhabit different environments. Conversely, the predicted probability of OD in crosses between 2 populations of the same species is low for populations with the same karyotype, isolated for <500 years, and that occupy similar environments. In the former case, we recommend crossing be avoided or tried on a limited, experimental basis. In the latter case, crossing can be carried out with low probability of OD. We used crosses with known results to test the decision tree and found that it correctly identified cases where OD occurred. Current concerns about OD in recently fragmented populations are almost certainly excessive.