Modelling RSD detection for sampling xylem and leaf-sheath biopsies

Detection of Leifsonia xyli subsp. xyli (Lxx) in planting material relies on sampling and pooling from a small number of stalks from a crop to make one sample. Multiple samples are taken depending on disease incidence and the value of the crop. In the Australian industry, two sample types are routinely used for diagnosing infection using molecular methods: xylem exudates and leaf-sheath biopsies. Each has their own advantages and disadvantages. Although both sample types are very good at detecting infection, the very nature of sampling means that there can never be a guarantee that samples returning a negative result equate to a disease-free crop. Baseline detection probabilities were obtained by sampling 100 stalks each from two varieties, using two sampling methods (xylem and LSB) and diagnosed using quantitative polymerase chain reaction (qPCR). Modelling was used to predict the number of pooled samples required to detect infection in crops with different disease incidence. The number of pooled samples required to detect Lxx for Q242A was similar for both sample types, while for Q208A the xylem sample required less pooled samples for detection. However, factors such as variety, recommended number of stalks pooled per sample and sampling effort required will determine which sampling method should be used. The detection probability work presented will assist the industry better understand RSD infections and the limitations of diagnosing it.