Emge, R.G. and Shrum, R.D. 1976. Epiphytology of Puccinia striiformis at five selected locations in Oregon during 1968 and 1969. Phytopathology 66:1406-1412.
Stripe rust lesion on wheat. Photo courtesy M. Nita, Kansas State University, Manhattan, KS.
Stripe or yellow rust is caused by Puccinia striiformis, an obligate parasite of wheat and other grass species. Infection is characterized by powdery masses (pustules) of yellow spores which form stripes on the leaf surface. The pustules have a characteristic torn margin and can occur on both sides of the leaves, stems, and heads. The pustules can cover large areas of affected leaves on young plants; on older plants the pustules are arranged in parallel lines. This arrangement gives affected leaves a characteristic striped appearance. Spores are mainly produced from the pustules on leaves and to a lesser degree sheaths and stems. P. striiformis can infect the glumes and awns at flowering resulting in an accumulation of spores in the florets and the surface of the developing grain; however, stripe rust is not seed-borne.
Conditions favoring establishment are wet, cool temperatures, approximately 8-15° C, although it can survive temperatures up to 40° C. Consequently, stripe rust is a serious problem on wheat in regions where cool temperatures prevail through the wheat growing season. It is especially prevalent in Europe and the Pacific Northwest of the USA. It is an occasional problem elsewhere, such as the US Great Plains, during early to mid-spring.
Disease progress during 1968 and 1969 in Oregon, in five locations differing climatologically, was studied by Emge and Shrum (1979). See their paper for specific details of the experiment. Experimental plots were 0.2 hectares (42.75 m x 45.75 m), planted with Omar wheat, a highly susceptible variety, and 25 g of urediospores/Ha were applied in the center of each treated plot. A plastic sheet was placed over inoculated plants at night to maintain favorable moisture levels. Sampling points were located along lines radiating from the inoculated center of the plot for a total of 40 samples per plot.
Stripe rust severity was estimated at weekly intervals. Since stripe rust pustules are discrete it is possible to calculate percent severity fairly easily. But in the areas where stripe rust expands, lesions often coalesce, and therefore, a code based on the size of the sporulating area and density of sporulation was used in a mathematical function to estimate disease severity. The AUDPC was then used as a summary of disease severity to compare the differences between the 1968 and 1969 growing seasons. The following R code provides a method to explore the AUDPC for 1968 and 1969, two years with different environmental conditions.
## Set up vector for Madras AUDPC ChartdaysAfterInoculation <- c(0,10,20,30,40,50,60,70,80,90,100)severityYear68 <- c(0,0,0,0,3,20,50,80, 90, 100, 100)severityYear69 <- c( 0,0,0,0,0,0,0,3,6,30,70)## Set up the line graph for 1968plot( daysAfterInoculation, severityYear68, type="o", pch=22, col="mediumblue", ylim=c(0,100), xlab='Days After Inoculation', ylab='% Infection')
## Set main titletitle(main="Madras Disease Progress")
## Overlay line for 1969 plotlines( daysAfterInoculation, severityYear69, type="o", col="orange")
## Set the legend of the graph in upper left cornerlegend( "topleft", c("1968","1969"), pch= c(22,21), lty=1, col = c("mediumblue","orange"), title="Year", inset=0.05)
## Calculate the AUDPC using the function called 'audpc' that## was created in the AUDPC exercise of this document## If you have started a new R session and did not save the## function you will need to create it again.## To check that you do have the function available, try## entering the function name, 'audpc', at the command## line and R should return the content of the function.
# Calculate the AUDPC for 1968audpc(severityYear68,daysAfterInoculation)
# Calculate the AUDPC for 1969audpc(severityYear69,daysAfterInoculation)
Does the AUDPC provide a good summary of the difference between these two disease progress curves? What information does it capture and what does it leave out?
The effects of disease resistance and fungicide applications on wheat leaf rust disease progress can be evaluated using the AUDPC. Typically resistance is the best way to control foliar diseases in wheat such as stripe rust and leaf rust (caused by Puccinia triticina). The following example, based on field data from Kansas (Sparks and Stack, unpublished), illustrates the difference in disease severity for a variety susceptible to leaf rust, Jagger, and the resistant cultivars 2137 and Cutter. The last sample date was June 3, after which the plants were senescing. Using the methods from previous examples, create a way to compare disease levels on the three varieties of wheat.
Timing is critical if a fungicide application is used to control wheat leaf rust. Typically the assumption can be made that the fungicide will be effective for only two weeks after application, preventing new infections during that time. Label restrictions prohibit the application of a fungicide after flowering has occurred. The critical protection period for wheat is the two weeks after flowering when the head is filling. Loss of photosynthetic area at this time will result in the greatest yield losses. Once this two week period has passed, losses to any new infections are much less severe.
Download this .csv file (right click -> 'Save Link As' or 'Save Target As') and import it into R using the read.csv command illustrated in "An Introduction to the R Programming Environment". Calculate the AUDPC using R and construct two graphs. One graph should include only the wheat with no fungicide treatment to compare the effects of resistance; the other should include both the non-treated and treated applications. How does the size of the fungicide effect compare to the effect of resistance? Would it be the same as resistance?
Since the fungicide typically is only active for two weeks after application, after looking at your graph, when do you think the fungicide was applied?
Next, performing linear regression in R.