Incorporating mate choice into conservation breeding programs can improve reproduction and the retention of natural behaviours. However, different types of genetic-based mate choice can have varied consequences for genetic diversity management. As a result, it is important to examine mechanisms of mate choice in captivity to assess its costs and benefits. Most research in this area has focused on experimental pairing trials, however this resource-intensive approach is not always feasible in captive settings and can interfere with other management constraints. We used generalised linear mixed models and permutation approaches to investigate overall breeding success in group-housed Tasmanian devils at three non-mutually exclusive mate choice hypotheses: (i) advantage of heterozygous individuals, (ii) advantage of dissimilar mates, and (iii) optimum genetic distance, using both 1,948 genome-wide SNPs and 12 MHC-linked microsatellites. The managed devil insurance population is the largest such breeding program in Australia and is known to have high variance in reproductive success. We found that non-genetic factors such as age were the best predictors of breeding success in a competitive breeding scenario, with younger females and older males being more successful. We found no evidence of mate choice under the hypotheses tested. Mate choice varies among species and across environments, so we advocate for more studies in realistic captive management contexts as experimental or wild studies may not apply. Conservation managers must weigh up the need to wait for adequate sample sizes to detect mate choice with the risk that genetic changes may occur during this time in captivity. Our study shows that examining and integrating mate choice into the captive management of species housed in realistic, semi-natural group-based contexts may be more difficult than previously considered.

"Farquharson et al EvolAppl metadata.xlsx"

This file comprises the following six sheets: a README with metadata to interpret the other sheets; 1) mate choice data: non-genetic and genetic information for each Tasmanian devil in the analysis, used to examine Hypothesis 1; 2) pedigree: the pedigree of offspring born in the free-range enclosures, used to test Hypotheses 2 & 3; 3) MHC genotypes: the raw genotypes at the 12 MHC-linked microsatellite loci, used to calculate standardised MHC-based heterozygosity and to test Hypotheses 2 & 3; 4) SNP genotypes: the genotypes at the 1948 SNP loci used to calculate standardised genome-wide heterozygosity; and 5) SNP data similarity: the genotypes at the 1948 SNP loci presented in two-column per loci format, used to test Hypotheses 2 & 3. Save each sheet separately as a .csv file for input directly into "Farquharson et al EvolAppl_code.R".

"Farquharson et al EvolAppl_code.R"

This R script can be used in conjunction with "Farquharson et al EvolAppl metadata.xlsx" to obtain the results of the various hypotheses tested. Save each sheet of "Farquharson et al EvolAppl metadata.xlsx" as a separate .csv file for input directly into the code. This code contains custom functions to calculate pairwise similarities of a group of breeding age individuals to examine observed pairwise similarities and compare these to pairwise similarities simulated under a random mating scenario in order to test mate choice hypotheses (e.g. advantage of dissimilar mates; optimum genetic distance).