Link to home

A Computer Program to Improve the Efficiency and Accuracy of Postulating Race-Specific Resistance Genes

May 2004 , Volume 88 , Number  5
Pages  545 - 549

Yeshi A. Wamishe , Department of Plant Pathology, University of Arkansas, Fayetteville 72701 ; Kevin C. Thompson , Agricultural Statistics Laboratory, University of Arkansas, Fayetteville 72701 ; and Eugene A. Milus , Department of Plant Pathology, University of Arkansas, Fayetteville 72701

Go to article:
Accepted for publication 5 January 2004.

Gene postulation has been the most widely used technique to determine the presence of particular rust resistance genes in lines of small grains. It applies the principles of gene-for-gene specificity to determine the most probable race-specific resistance genes present in host lines. As the numbers of lines, resistance genes, and races increase, postulation based on visual comparisons of infection types becomes more complex and laborious, and errors may occur. A computer program was developed to facilitate identification of race-specific leaf rust (Lr) genes in wheat (Triticum aestivum). Seedlings of 116 contemporary lines of soft red winter wheat and 24 Thatcher isolines (each Thatcher isoline with a single Lr gene) were inoculated with 22 races of Puccinia triticina. Infection types were recorded on the standard 0 to 4 scale where infection types 3 and 4 were considered high (line was susceptible; race was virulent) and others were low (line was resistant; race was avirulent). Based on the gene-for-gene concept, lines susceptible to a particular race cannot have an Lr gene for which the race is avirulent. For each line, step 1 of the program summarized results from races that were virulent on the line to definitively exclude Lr genes from the line, and this exclusion resulted in a relatively short list of Lr genes that could be present. Step 2 of the program utilized data from races that were avirulent on the line, and the output listed the low infection types produced on the line and the isolines with Lr genes that were not excluded in step 1. Of these Lr genes, a gene was considered present if the low infection type produced on the line by one or more races matched the low infection type on the corresponding isoline. Otherwise, the gene was considered possibly present. Epistatic effects of one or more Lr genes prevented definitive inclusion or exclusion of genes considered possibly present. If the low infection type produced on the line was lower than that on any of the isolines listed in step 2, then the line was considered to have an unidentified Lr gene; i.e., a gene that was not in the set of 24 isolines. This program facilitated the objective and accurate postulation of Lr genes and could be adapted to other host-pathogen systems.

© 2004 The American Phytopathological Society