Submitted to: Journal of Virological Methods
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: March 23, 2010
Publication Date: August 20, 2010
Repository URL: http://handle.nal.usda.gov/10113/62130
Citation: Grisham, M.P., Johnson, R.M., Zimba, P.V. 2010. Detecting Sugarcane yellow leaf virus infection in asymptomatic leaves with hyperspectral remote sensing and associated leaf pigment changes. Journal of Virological Methods. 167(2):140-145. Interpretive Summary: Sugarcane yellow leaf in sugarcane is caused by a virus that infects the sugarcane plant, but rarely causes visible symptoms until late in the growing season. Even though visible symptoms may not develop before the infected sugarcane is harvested, yield may be reduced. It is also important to know if stalks are infected with the virus when they are cut for planting which typically occurs 3 to 4 months before symptom expression. Without visual symptoms, a laboratory test is currently needed to determine if plants are infected with the virus. An experiment was conducted to determine if light reflected from leaves infected with the sugarcane yellow leaf virus was different from light reflected from healthy leaves. In the experiment, reflected light was measured from leaves of virus infected and healthy plants of two sugarcane varieties. The light reflected from leaves of the test plants was analyzed with an instrument designed to measure the amount of different colors of light including ultraviolet, visible colors (violet, blue, green, yellow, orange, and red), and near infrared. Although no visible symptoms of sugarcane yellow leaf were observed among the test plants, differences in the light reflected by healthy and virus infected were observed. Using information collected, one could correctly predict virus-infected plants between 86 and 89% of the time depending on the variety. The concentration of different pigments in leaves affects the reflectance of light from leaves, each pigment absorbing a different part of the light spectrum while reflecting the rest. The concentration of two pigments was reduced in leaves of virus-infected sugarcane plants. Developing a technology based on leaf reflectance to determine if sugarcane plants are infected with the yellow leaf virus would provide a more efficient method of detecting diseased plants without the need to collect plant tissues for laboratory diagnostic tests.
Technical Abstract: Sugarcane yellow leaf caused by Sugarcane yellow leaf virus (ScYLV) does not produce visual symptoms in most susceptible plants until late in the growing season. An experiment was conducted to determine if leaf reflectance and pigment analysis could be used to determine ScYLV infection prior to symptom development. High-resolution, hyperspectral reflectance data from leaves of two cultivars, LCP 85-384 and Ho 95-988, were measured and analyzed on three dates, 13 July, 12 October, and 4 November 2005. Approximately one half of the sampled leaves of each cultivar was infected with ScYLV as determined by reverse transcriptase-polymerase chain reaction (RT-PCR) analysis. Leaf pigments were extracted from leaf samples collected on 12 October and analyzed for chlorophylls and carotenoids concentrations. Results from discriminant analysis showed that leaf reflectance was effective at predicting ScYLV infection in 64% and 57% of the cases if varieties were combined using resubstitution and cross validation techniques, respectively. Predictions of plants not infected with ScYLV were correct in 72% and 65% of the cases using resubstitution and cross validation techniques, respectively. The predictive value of leaf reflectance improved if cultivars were analyzed separately; ScYLV infection was correctly predicted in 73% of the cases in both cultivars using both resubstitution and 63% and 62% of LCP 85-384 and Ho 95-988, respectively, using cross validation. Correct prediction of plants without ScYLV was also improved when results from cultivars were analyzed separately. The most distinct differences between the reflectance of leaves from ScYLV-infected and control plants were observed in several spectral regions among samples collected on 4 November. ScYLV infection influenced levels of several of the plant pigments studied including significant pigment concentrations in violaxanthin and '-carotene in the samples that were infected with ScYLV. Pigment data was effective at predicting ScYLV infection in 80% of the samples in the combined data set using the derived discriminant function with resubstitution, and 71 % with cross validation. Developing technology to remotely detect ScYLV infections without a laboratory based diagnostic technique would provide an efficient method to insure growers that the seed cane they are planting is free of the ScYLV. Monitoring for ScYLV infection when increasing seed is particularly critical because once infected the only reported method of freeing infected seed cane of ScYLV infection is by meristem tip culture.