Submitted to: American Journal of Potato Research
Publication Type: Peer reviewed journal
Publication Acceptance Date: 5/7/2003
Publication Date: 1/7/2004
Citation: Haynes, K.G., Weingartner, D.P. 2004. The use of area under the disease progress curve to assess resistance to late blight in potato germplasm. American Journal of Potato Research 81:137-141. Interpretive Summary: Potato varieties differ in their susceptibility to late blight, a serious disease of potato. Usually multiple observations on the severity of the disease are taken during the course of a late blight epidemic and those observations are subsequently used to calculate the area under the disease progress curve. Using data collected from a late blight epidemic on potatoes in Florida in 1997, in which disease was estimated 14 times during the 31-day epidemic for 23 different potato varieties, we showed that differences in the reaction of potato varieties to late blight could be accurately determined using data collected at two distinct phases of the late blight epidemic. For accurately comparing the reaction of potato varieties to late blight, disease needs to be estimated only once early in the epidemic (within the first two weeks in this data set) and only once as the epidemic is reaching its peak (between 14 and 21 days in this data set). This results in a considerable savings in time and cost. This information will be of interest to plant geneticists, breeders, and pathologists involved in developing disease resistant varieties in several crops.
Technical Abstract: The multiple evaluation of potato cultivars and breeding selections (clones) for disease during the season can be costly and may not be necessary for accurate assessments of disease resistance or susceptiblity. For diseases whose progression can be described by sigmoid curves, an estimate of the area under the disease progress curve from two data points may provide as much information as from repeated assessments. Twenty-three clones were planted in a randomized complete block design in Hastings, Florida in 1997 and evaluated for percent diseased foliage 14 times during a 31-day period after late blight was first noticed in the plots. The area under the disease progress curve (AUDPC) was calculated for all 14 assessments. The AUDPC was also calculated for nine sub-sets of the original data set. There was a high correlation between all the methods in the calculation of the AUDPC. There was also a high correlation between the rankings of the clones for AUDPC as calculated for these various sub-sets. The AUDPC calculated based on two dates (representing the beginning of the epidemic and the time until two of the clones were dead from late blight) was as informative as the AUDPC calculated on the entire data set. The AUDPC could be calculated based on any two dates from this time frame as long as one date was shortly after the epidemic started and the other was as the epidemic was reaching its peak. Much information was lost if AUDPC was calculated based on dates involved only in the early part of the epidemic. A considerable savings in time and effort can be realized by only a few assessments.