Assessing rarity: genomic insights for population assessments and conservation of the most poorly known Amazonian trees

Tropical forests comprise a few hyperdominant and many rare tree species, but distinguishing the truly rare from those under-sampled remains a challenge for ecology and conservation. Given the vastness of Amazonia (~6 million km2, ~3.9x1011 individual trees), increasing sampling cannot solve this problem. Still, half of all species are known from three or fewer collections, making predicting their abundances and distributions impossible with census data alone. Here, we integrate census data with next-generation genomics to assess the rarity of one of the most poorly known and highly threatened Amazonian trees, Magnolia yantzazana. Genetic analyses indicate that while there is relatively high nucleotide diversity among sequences (π > 0.5), there is also evidence of a loss of heterozygosity (He > Ho) and inbreeding (FIS > 0.5), consistent with a small, isolated population. Demographic reconstructions show population decline since the late Pleistocene, with a predicted effective population size (Ne) of ~103 in recent millennia. Together, the low heterozygosity, potential inbreeding, demographic trajectory, and census data suggest M. yantzazana is in fact a truly rare species, highly vulnerable to ongoing environmental change and anthropogenic threats in the region, notably mining, and support updating its conservation status to Critically Endangered (CR). This study offers a framework for using genomic tools to advance our understanding of the rarest Amazonian trees and establishing conservation priorities, despite the limited field collections available for most species.