Phytophthora species are globally significant soilborne oomycetes responsible for widespread ecosystem decline. Standard soil sampling protocols, originally developed for qualitative baiting assays, typically require collecting substantial soil volumes in order to capture viable propagules. While effective for culture-based detection, these protocols are labour-intensive and can damage the shallow root systems of sensitive host species such as New Zealand kauri (Agathis australis). Phytophthora agathidicida (PA), the pathogen associated with kauri dieback disease, is routinely surveyed using these methods. However, quantitative data describing the vertical distribution of PA in natural forest soils are lacking. Consequently, it remains unclear whether extensive depth sampling is necessary to ensure consistent molecular detection. In this study, we applied a quantitative oospore DNA (oDNA) qPCR assay to characterise the fine-scale vertical distribution of PA across four soil depth increments (0-5, 5-10, 10-15, 15-20 cm) from 12 kauri trees representing a range of disease stages. Results revealed distinct vertical stratification, with PA DNA concentrations peaking within the upper 0-10 cm of soil in non-symptomatic and early-stage trees. In later-stage symptomatic trees, the absolute peak occasionally reached 10-15 cm, while pathogen signals remained consistently detectable within the top 10 cm. Field validation from an additional eight trees confirmed that targeted 0-10 cm "shallow" sampling yielded higher PA concentrations than deeper sampling protocols. These findings provide a data-driven basis for refining soil sampling strategies, enabling more sensitive molecular detection while minimising disturbance and logistical effort in fragile ecosystems.
Depression subtype classification from social media posts: few-shot prompting vs. fine-tuning of large language models
BackgroundSocial media provides timely proxy signals of mental health, but reliable tweet-level classification of depression subtypes remains challenging due to short, noisy text, overlapping symptomatology,



