Journal Article
Genetic diversity and population structure of Myrciaria dubia from the Peruvian Amazon: implications for germplasm conservation and crop improvement
Eduardo Mejia de Loayza; Guillain Estivals; Diana Castro-Ruiz; Werner Chota-Macuyama; Carlos Angulo-Chavez; Mike Corazon-Guivin; Ángel Rodriguez del Castillo; Jhon Alvarado Reategui; Carlos Angulo-Villacorta; Kember Mejia; Dennis Del Castillo Torres; Carmen García-Dávila
Genetic Resources and Crop Evolution · Vol. 73, Issue 4 · 2026
Abstract
Myrciaria dubia (camu-camu) is a shrubby fruit tree native to the continental Amazon whose fruits have been intensively harvested from wild stands, potentially reducing effective population sizes. We quantified genetic diversity and population structure across seven wild Peruvian Amazon populations and delineated river-basin genetic units to guide provenance-aware germplasm conservation and breeding. We genotyped 254 individuals from the Napo, Ucayali, Nanay, Tahuayo, Putumayo, Tigre, and Curaray basins using six polymorphic microsatellite loci. Overall, 48 alleles were detected. Observed heterozygosity (0.149–0.483) was generally lower than expected heterozygosity (0.220–0.531), and population-level inbreeding coefficients (F IS = − 0.038–0.560) indicated significant heterozygote deficits in Napo, Curaray, and Tahuayo. The Putumayo population harbored nine private alleles, representing a unique genetic reservoir. Pairwise differentiation was substantial (F ST = 0.093–0.660; Nei’s distance = 0.068–1.734), with the strongest divergence between Tigre and Ucayali. Neighbor-joining, Bayesian assignment, and Discriminant Analysis of Principal Components (DAPC) initially supported three major genetic units and highlighted Putumayo as genetically isolated; additionally, hierarchical STRUCTURE analyses resolved eight clusters, and DAPC distinguished seven population-specific groups. Analysis of molecular variance attributed 56.5% of the variation within individuals and 34.7% among populations. Mantel tests supported isolation by distance based on straight-line geographic distances ( r = 0.53–0.56; p ≤ 0.017), whereas river-network distances were not significant. Overall, the data indicate a geographically structured genetic architecture shaped by dispersal limitation and basin-scale differentiation, supporting three provenance units for germplasm banking and breeding: (i) Napo–Ucayali–Nanay–Tahuayo, (ii) Tigre–Curaray, and (iii) Putumayo.