Back

Conservation genomics of desert tortoises (Gopherus agassizii) in the Colorado Desert

Jennings, W. B.; Lovich, J. E.; Cummings, K. L.; Puffer, S. R.; Ennen, J. R.; Agha, M.; Brundige, K. D.; Tracy, C. R.

2025-12-10 genomics
10.64898/2025.12.08.692717 bioRxiv
Show abstract

Management units (MUs) are important to conserving species of conservation concern. Although the MU definition is simple (i.e., a demographically independent population), identifying MUs in practice is difficult because investigators must choose a recent (i.e., last few generations) migration rate threshold. One suggested MU criterion is m < 0.10, where m is the proportion of a subpopulation comprised of recent migrants. However, a more recent study found that m as low as 0.02 (i.e., one migrant per generation) can cause estimates of linkage disequilibrium effective population size (NeLD) to underestimate a subpopulations true Ne. Here, we used genome-wide SNP data obtained from 40 tortoises and an MU criterion of m < 0.02 to define MUs within the Colorado Desert population of the Mojave desert tortoise (Gopherus agassizii)--an endangered species in California and a species in the USA protected under the Endangered Species Act. Based on our results we defined the Mesa and Deep Canyon subpopulations as MUs even though the latter received at least two recent immigrants. Migrant(s) from the Shavers Valley subpopulation were likely translocated by humans, whereas migrant(s) from the Mesa subpopulation could have been human-mediated or natural migrants. Initial analyses suggest that the Shavers Valley subpopulation may be an MU, but our low sample size renders this result inconclusive. The local Ne estimates for the Mesa and Deep Canyon subpopulations--5 and 13 adults, respectively--are below the minimum sizes according to the 50/500 rule and thus they may be candidates for genetic rescue.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.