"One
picture is worth a thousand words."
--Foreword
for the book "GGE Biplot Analysis: a Graphical Tool for Breeders,
Geneticists, and Agronomists"
By
Dr. D. E. Falk
One of the most
well-developed and acute faculties of the human brain is the ability to
perceive, analyze, and interpret complex visual information.
Patterns, or visual relationships, are the easiest to understand
quickly. Comprehension of visual
information is much better and faster than linear or tabular numerical
information.
Extensive data
bases of complex information are being meticulously accumulated and maintained
in many research programs, but they are not being utilized to their full
potential because interpretation of them is too complicated.
The principal component methodology
and the GGEBiplot software described in this book can change that by making
interpretation of complex data sets as easy as a few key strokes and a few
minutes of viewing easy-to-understand graphic displays.
Using GGEBiplot
software requires minimal preparation of data--simple two-way data spread
sheets are the basic input format. The
initial graphic output can be easily modified from drop-down menus.
The data set being analyzed can be manipulated by removing various
parameters and entries during the real-time massaging of the data set and
different statistical analyses can be performed.
Several alternative ways of viewing the output are possible to assist
in interpretation of the relationships among the parameters, among the
entries, and their interactions, and it is all done
'live'
in real- time. Information can be
analyzed faster, and can be understood more completely, using the GGEBiplot
software visual display than the data tables and graphic displays generated by
most spreadsheets and database systems. The
graphic output can be printed to produce a hardcopy version, and/or saved as
an image and used in presentations or publications.
A log file of the relevant statistics and correlation matrices is also
produced and automatically saved for later detailed reference.
The GGEBiplot
can help any one learn more about the interactions among a multitude of
traits/parameters in their own data than ever before.
GGE software has tremendous potential value to plant breeders,
agronomists, pathologists, physiologists, nutritionists, and anyone working in
an applied science field. It is
currently being used in the cereal breeding program at the University of
Guelph to evaluate overall agronomic merit, quality, genotype x environment
interaction for numerous traits, genotype x trait interactions, and trait x
environment interactions in breeding lines being advanced through the testing
system, to select parents and parental combinations for crossing, to evaluate
relationships among traits (especially quality), to identify determinants of
yield and quality factors in the populations, and to assess the discriminating
value and stability of various testing locations.
GGE Biplots have
been utilized in chromosome mapping of morphological and molecular markers,
and to identify and locate QTL markers in doubled haploid barley mapping
populations. The biplot concept
is also used to determine the physiological associations of morphological
traits and kernel characteristics with quality factors.
This is especially valuable when both negative and positive
correlations exist among several desired traits.
Biplots can also be used to identify and quantify the effects of
environment on one or more traits across a range of genotypes, and the effects
the environment may be having on the relationships among those traits.
This is particularly useful when selecting among elite lines to advance
to more extensive and intensive trials, and in choosing parents for the next
cycle of improvement in complex traits with interactions among the components.
Biplots have
been used in evaluating students' performance in a course to determine which
assignments and exercises are most likely to contribute to their understanding
of concepts, based on association with performance on later examinations.
Essentially, any
two-way data set with multiple entries and multiple parameters can be analyzed
using GGEBiplot; the result will be a better understanding of the interactions
among entries, among parameters, and interactions between entries and
parameters. The GGEBiplot
software offers convenience, the resolving power of principal component
analysis, speed, and simplicity of interpretation.
The potential applications of the principal component analysis
technique using this amazingly friendly and informative software are limited
only by the imagination of the researcher. GGEBiplot is simple and fun to use;
it is much like a useful and productive video game for scientists.
Duane
E. Falk, PhD
Associate
Professor and Cereal Breeder
Department
of Plant Agriculture
University
of Guelph
Guelph,
Ontario
Canada
Adjunct
Associate Professor
Department
of Plant Biology
Faculty
of Natural and Agricultural Sciences
University
of Western Australia
Crawley,
Western Australia
Australia
