Unlocking species and subspecies diversity of harmful microalgae using HiFi long-read metabarcoding: insights from...

Accurate species-level identification is essential for tracing toxin sources, understanding bloom dynamics, and improving monitoring. However, short-read metabarcoding often lacks resolution for closely related harmful microalgae. We applied PacBio HiFi long-read metabarcoding of ITS1-5.8S-ITS2 and partial LSU (D1-D2) rDNA to 600 eDNA samples from the China Sea, spanning ∼40° of latitude (2019-2024). Using a reproducible QIIME 2 workflow, a curated Amphidomataceae reference set, and species-resolution criteria (p-distance= 0.040 in Azadinium; 0.045 in Amphidoma), we resolved 44 Amphidomataceae lineages (17 named species, 27 putative new taxa) and mapped their fine-scale biogeography. The four azaspiracids (AZAs) producers exhibited distinct distributions: Az. dexteroporum occurred along warm saline shelves of the East and South China seas; Am. languida was restricted to temperate nearshore areas in the Yellow Sea and Yangtze estuary; Az. poporum was observed mainly nearshore with ribotype segregation (B and C widespread); and Az. spinosum was widespread in the East and South China seas, dominated by ribotype A and B. Niche metrics and clustering of sea-surface temperature (SST), salinity, and trophic index (TRIX) defined three ecological groups consistent with MaxEnt habitat suitability. Likely AZAs sources are Am. languida and Az. poporum in the Bohai/Yellow seas, and Az. dexteroporum, Az. poporum, and Az. spinosum in the East/South China seas. HiFi long-read metabarcoding resolves hidden diversity, strengthens risk assessment, and provides a reproducible, scalable framework for global species- and subspecies-level mapping of harmful microalgae.