Frequent hybridization and polyploidization in the plant kingdom often blur species boundaries, posing a major challenge for traditional morphology-based identification methods.

In a study published in Annals of Botany, researchers from Xishuangbanna Tropical Botanical Garden (XTBG) of the Chinese Academy of Sciences have developed a new framework that integrates stable trait analysis, herbarium re-examination, and computational modeling to accurately identify species within complex polyploid groups. The new framework reveals that up to 50% of historical herbarium specimens may have been misidentified, offering critical insights that could reshape conservation strategies for polyploid plants.

Using the East Asian polyploid complex Rorippa dubia–indica (a group of tetraploid and hexaploid plants in the mustard family) as a model system, the researchers analyzed over 5,000 specimens, including 3,136 fieldcollected samples and 2,015 herbarium records dating from 1893 to 2021.

By applying computational modeling to reconstruct the decisionmaking logic behind specimen identification, the researchers found that 12–50% of both historical and modern specimens had been misidentified. These errors were largely driven by reliance on “plastic traits” (e.g., leaf shape and fruit length), which vary with environmental conditions, rather than on stable, evolutionarily conserved characteristics. They identified three reliable diagnostic traits: seed arrangement, number of petals, and genome size.

When the researchers corrected species labels and rebuilt species distribution models (SDMs), the projected habitats shifted dramatically. For example, Rorippa indica showed significant niche loss across central and southern China under future climate scenarios, a trend that was completely obscured by the original, errorprone identifications.

“Our framework demonstrates the importance of integrating morphological and phylogenetic inference with machine learning tools to resolve taxonomically difficult polyploid complexes. This approach provides a transparent, replicable pipeline for secondary specimen evaluation,” said XING Yaowu of XTBG.

The researchers also call for systematic reidentification of taxonomically challenging groups, especially those with medicinal, agricultural, or ecological importance, to generate more accurate and reliable species data, thereby supporting biodiversity conservation and research.

Published: 05 March 2026