Establishment Of Dynamic Prediction Model Of Rubber Tree Anthracnose Based On Real-time PCR Detection Technology

Rubber tree anthracnose caused by Colletotrichum spp.is an important reason for the decline of rubber production in Asia.Colletotrichum siamense shows high environmental fitness and pathogenicity,and is the main pathogen of rubber anthracnose in China.In this study,a real-time PCR detection technology was established for the detection of C.siamense.The effects of environmental factors and the phenology of rubber tree on the growth and pathogenic processes of C.siamense in rubber trees were studied.Finally,the dynamic prediction model of rubber tree anthracnose was preliminary established.These results provided technical support for the prediction and control of the disease.The results are as follows:1.A pair of real-time PCR primers ITSF3/ITSR3 with good specificity were screened out using the ITS sequence of C.siamense as the template.A real-time PCR system was established,the system has high sensitivity and strong stability,which could detect 100 fg genomic DNA,100 copies of target DNA and 20 conidia.The system was not affected by the DNA extract of rubber trees,and the C.siamense could still be detected in 0.0001%of infected tissues.The real-time PCR system can monitor the growth process of C.siamense from latent infection to lethal infection in rubber leaves at different phenological stages.2.The effects of environmental factors and phenological phase on the growth and pathogenic processes of C.siamense were studied.The results of laboratory simulation of the isolated leaves showed that,under the conditions of 100%relative humidity and28℃,C.siamense had the fastest rate of infection on isolated leaves.The younger the leaves,the shorter the incubation time and the faster the infection rate.3.A dynamic prediction model of rubber anthracnose was established based on real-time PCR based on the environmental factors,phenological phase,progression of disease and quantitative data of C.siamense:Y=2^E-05[X₄(-0.211+0.05 X₂-0.157X₃+0.796 Ln(10^-5X₅²+0.0121X₅+0.984)-0.427Ln X₄+0.863Ln（（0.065（X₁-23.435））²sin（0.155（X₁-23.435））+1.578）)]^5.616%,R² value was 0.809.A binary Logistic regression equation was established to verify the dynamic prediction model:Logit（P）=-56.380+0.174X1+0.443X2-1.666X3+0.107X4+0.288X5,Hosmer-Lemshaw test showed that the significance of the model is 0.969,showing good fitting effect.In the model,Y represents the disease index,P represents the probability of disease occurrence,and X1-X5 represent the temperature,humidity,phenology,fungal quantity and time.4.The accuracy rate of the dynamic prediction model verified by field data is 83%,and the Logistic regression model can accurately exclude false positives and improve the accuracy of the model.Compared with other early detection method,real-time PCR method established in this study shows highly specificity and sensitivity and could quantify the amount of C.siamense in rubber tree leaves in real-time.In addition,the laboratory prediction model of rubber anthracnose was established based on real-time PCR technology could be used for the dynamic prediction of rubber tree anthracnose.