Biomathematics & Statistics Scotland
Many plant diseases, such as root rot in raspberries, are assessed visually at regular intervals using ordered categories such as the five point scale from 1 (healthy) to 5 (dead). These scores have a degree of subjectivity and are made by different assessors on different dates. Principal co-ordinates analysis can combine information across dates and assessors. Pairwise similarities, based on all the disease assessments, are derived for all plants. These similarities are then condensed into a small number of dimensions or co-ordinates.
In a raspberry root rot trial two cultivars and their offspring were scored on nine dates over a two-year period by three assessors. The first principal co-ordinate explained 43% of the similarity matrix and showed a strong spatial effect related to the slope of the field and consequent changes on soil moisture content. Heritabilities were calculated after removing this spatial effect and found to be significant for the first and third principal co-ordinates. By combining these principal co-ordinates with genetic marker data, quantitative trait loci for root rot resistance have been identified.
Further details from: Christine Hackett
Article date 2007
