Molecular identification of protected Magnoliaceae species: a multi-marker evaluation based on extensive sampling and genome-skimming data

Magnoliaceae are among the most threatened plant families, with 24 species listed as nationally protected in China. The reliability of DNA barcoding for their identification has not been systematically evaluated under large-scale individual sampling. Here we employed genome skimming to generate plastid genome and nuclear ribosomal internal transcribed spacer sequences from 102 individuals representing 23 protected Magnoliaceae species. Supplemented with publicly available data, we assessed DNA barcoding performance at both species and individual levels across 196 chloroplast fragments using three analytical methods. At the species level, the BLAST-based method resolved 18 of the 24 species using ten selected candidate fragments, considerably more than standard barcodes or even complete plastid genomes could achieve under the same method. Under tree-based and barcode gap-based methods, however, complete plastid genomes yielded the highest resolution. When comparing the same fragments individually, species-level success rates were generally lower than individual-level success rates owing to the stringent criterion requiring all individuals of a species to be correctly identified. We term this systematic divergence the Species-Individual Discrepancy, which demonstrates that species-level resolution does not guarantee individual-level identification reliability. Nine species could not be correctly identified by any single fragment, likely due to incomplete lineage sorting or chloroplast capture. Our findings show that intraspecific polymorphism and reticulate evolution complicate plastid-based barcoding. We therefore recommend integrating population-level reference data and multi-marker strategies in conservation applications for threatened Magnoliaceae species.