Investigating Genetic differences between populations of Ciborinia camelliae collected from different locations.
by
Ruth Butler
Crop & Food Research, Lincoln, Private bag 4704, Christchurch, New Zealand
Coauthors: Gillian Arnold (Long Ashton Research Station, Department of Agricultural Sciences, University of Bristol, Long Ashton, Bristol, UK BS41 9AF), Ron van Toor, Hayley Ridgway, Alison Stewart (Soil Plant and Ecological Sciences Division, PO Box 84, Lincoln University, Lincoln, Canterbury, New Zealand)

Camellia blight (Ciborinia camellia) is a fungal disease of camellia bushes that causes the flowers to become prematurely brown and rotten. The disease has recently become widespread in New Zealand, having been prevalent in the USA and Japan for several years. It forms hard fruiting-bodies (sclerotia) that remain in the soil beneath camellia bushes and then infect the flowers as they form, by firing ascospores from an apothecium extended from the sclerotium. In nature, the blight is caused mainly by the dispersion of ascospores spread on the wind, but in practice, the major cause of dispersion is through soil transported with bushes being distributed by commercial growers and gardeners. Methods for controlling this disease are currently being investigated (van Toor, 2002). Within this study, the genetic diversity of Ciborinia camelliae populations was examined, as the effectiveness of control measures may be affected by a lack of homogeneity between the populations.

Ciborinia camellia sclerotia were collected from three regions of the USA and four regions of New Zealand, with samples taken from three to twelve locations within each region. Material from each location was examined with universally primed polymerase chain reaction (UP-PCR, Bulat Mironenko, 1996) using six primers, which generated presence/absence information for 63 unique banding patterns from the genome of each location.

The standard method for exploring differences and similarities for this type of data has been cluster analysis, looking for groupings amongst the collections in the dendrogram obtained from the analysis. However, this approach ignores the known groupings (regions) from which the samples have been taken. We have used canonical variates analysis (CVA) as an alternative approach for exploring the genetic differences between collections from different regions. We present the results using canonical variates bi-plots, such as those suggested by Gower Hand (1996) and which have been partially implemented in SAS (e.g., Friendly, 2000). We discuss problems with the approach for these types of data, as well as advantages and disadvantages of CVA in comparison to the standard cluster analysis method.

References

Bulat S. A. and Mironenko, N. V. (1996). Identification of fungi and analysis of their genetic diversity by polymerase chain reaction (PCR) with gene-specific and non-specific primers. Russian Journal of Genetics 32 (2): 165-183.

Friendly, M. (2000). canplot.sas: Canonical discriminant structure plot. Available via the Internet at http://www.math.yorku.ca/SCS/sasmac/canplot.html

Gower, J.C. and Hand, D.J. (1996). Biplots. Chapman Hall

van Toor, R.F. (2002). Development of biocontrol methods for camellia flower blight caused by Ciborinia camelliae. PhD thesis, Lincoln University, Lincoln, New Zealand (unpubl).

Date received: August 29, 2002