Anthropogenic disturbances have long been acknowledged to be one of the primary threats to biodiversity worldwide, while little is still understood about how human-built infrastructure affects the gene flows and phylogeographic structure of plants. Such information is helpful for the conservation and restoration of human-disturbed ecosystems. Here we studied effects of a big river with a huge reservoir and two expressways on an island-endemic Primulina heterotricha (Gesneriaceae) on Hainan Island, China, one of the key parts of the glob ally important Indo-Burma biodiversity hotspot. By applying comparative phylogeography using one nuclear ribosomal DNA and two chloroplast DNA sequences, we estimated the levels of genetic diversity and differentiation in 176 and 117 individuals collected, respectively, before (in 2016) and after (in 2022) the construction of two expressways in Hainan Island, from the same eight populations of P. heterotricha. We found that Primulina heterotricha significantly increased nuclear genetic differentiation during the period 2016–2022, which coincides with the opening of the two expressways. Also notably, the sharing of ribotypes among the three groups of populations separated by the expressway network almost disappeared for the same period. Moreover, the changes in the significance of genetic barriers before and after road construction suggest that geographic isolation caused by both the reservoir and the expressways is key for understanding the present phylogeographical patterns of P. heterotricha. We provide direct evidence that anthropogenic infrastructures such as reservoirs and expressways have been capable of increasing genetic differentiation and, thus, modifying the phylogeographical pattern of P. heterotricha, in just a six-year period (or two generations of the study plant). To mitigate such negative pressure, we suggest establishing ecological corridors to enhance gene exchange between the two sides of the anthropogenic barriers.

Increased access to genome-wide data provides new opportunities for plant conservation. However, information on neutral genetic diversity in a small number of marker loci can still be valuable because genomic data are not available to most rare plant species. In the hope of bridging the gap between conservation science and practice, we outline how conservation practitioners can more efficiently employ population genetic information in plant conservation. We first review the current knowledge about the within-population genetic variation and among-population differentiation in neutral genetic variation (NGV) and adaptive genetic variation (AGV) in seed plants. We then introduce the estimates of among-population genetic differentiation in quantitative traits (QST) and neutral markers (FST) to plant biology and summarize conservation applications derived from QST–FST comparisons, particularly on how to capture most AGV and NGV on both in-situ and ex-situ programs. Based on a review of published studies, we found that, on average, two and four populations would be needed for woody perennials (n = 18) to capture 99% of neutral and adaptive genetic variation, respectively, whereas four populations would be needed in case of herbaceous perennials (n = 14). On average, QST is about 3.6, 1.5, and 1.1 times greater than FST in woody plants, annuals, and herbaceous perennials, respectively. We suggest using maximum QST rather than average QST among trait comparisons. Hence, conservation and management policies or suggestions based solely on inference on FST could be misleading, particularly in woody species. We recommend conservation managers and practitioners consider this when formulating further conservation and restoration plans for plant species, and for woody species in particular.