HORTSCIENCE VOL. 41(1) FEBRUARY 2006 30 The Global Need for Plant Breeding Capacity: What Roles for the Public and Private Sectors? Michael Morris Environmentally Sustainable Social Development Division, World Bank, Washington, DC 20433 Greg Edmeades Maizways, 43 Hemans Street, Cambridge, New Zealand Eija Pehu Agriculture and Rural Development, World Bank, Washington, DC 20433 For several decades after plant breeding emerged as a recognized field of science in the late 19th century, almost all plant breeding activities took place in public institutes, and al- most all plant breeders received their scientific education in public universities. Over time, the locus of plant breeding gradually shifted to the private sector, driven by the commercialization of agriculture and the associated privatization of agricultural research. The training of plant breeders, however, remained largely a public undertaking. Then, as now, private firms had few incentives to invest in an activity whose benefits accrue over a long period and are dif- ficult to appropriate, since plant breeding skills are not firm-specific and breeders can easily offer their services to rival employers. For more than 100 years, the arrangement worked well even as it continued to evolve. The training of plant breeders, recognized as being a public good, was paid for mainly by govern- ments, and plant breeding research, considered increasingly to be a private good, was funded more and more by private firms, mainly seed companies. Today, however, the arrangement is threatening to unravel. The numbers of students entering universities to be trained as plant breeders appear to be falling, as do the numbers of university graduates available to the plant breeding profession. Challenged to attract qualified staff, universities and seed companies in industrialized countries have been filling positions by hiring experienced breeders away from positions in developing countries. In recent years, even this strategy has not been enough: the number of new entrants into the plant breeding industry has continued to decline, and the plant breeding profession has grayed noticeably, especially in the public sector. Many fear that if current trends are not reversed, the plant breeding industry will soon face a critical shortage of skilled breeders. This paper examines factors that have contributed to the decline in capacity building within the agricultural sciences and considers changes that will be needed to prevent what some feel could develop into a crisis for the international plant breeding industry. In addi- tion to this introduction, the paper includes four sections. Section 2 discusses the knowledge and skills needed by plant breeders. Section 3 discusses how that knowledge and those skills have been acquired in the past and how they are being acquired today. Section 4 summarizes trends affecting global agriculture and shows how these have contributed to declining invest- ment in the education of agricultural scientists. Section 5 identifies actions that will be required to rebuild capacity in international plant breed- ing and contemplates needed changes in the roles of key actors. PLANT BREEDERS AND PLANT BREEDING What do plant breeders do? Observed plant performance or phenotype (P) can be described algebraically as P = G + E + (G × E) + e, where G = effects arising from the plant’s genetic con- stitution or genotype, E = effects attributable to the environment, G × E = effects that reflect the interaction of G with E, and e = effects due to random error. Plant breeders are concerned with dissecting P into its various components. The purpose of the dissection is to be able to manipulate G effectively and efficiently, so that genetic gain can be maximized. Plant breeding is both an art and a science, where the art encompasses careful observations of plant behavior in the field and to some extent the choice of parents for crosses, and the science relates to knowledge of genetics, physiology, pathology, entomology, statistics, and other disciplines (Lamkey, 2003). Skills needed by plant breeders. The cen- tral skill in plant breeding has always been knowledge of genetics, the study of the par- ticulate nature of inheritance, where the units of inheritance are individual genes. Modern quantitative genetics as learned by today’s students is based on statistical models, but the underlying theory continues to be grounded in the elementary laws of inheritance formulated by Mendel nearly 150 years ago. Simple traits such as flowering date typically have a high heritability (i.e., a large proportion of the ob- served variation for these traits is also observed in the next generation) and are regulated by few genes with large effects. Complex quan- titative traits such as yield have a relatively low heritability (i.e., they are controlled by many genes that interact with each other and with the environment, sometimes unpredict- ably). Plant breeders use their knowledge of quantitative genetics, with its foundation in statistics, mathematics, and Mendel’s laws, to estimate heritability, genetic variation, gene action, and G × E interaction—all critical to the design of an efficient breeding and testing system. Most successful breeders also have a good functional knowledge of physiology, pathology, entomology, soil science, and experimental design. Until the late 1980s, the plant breeder’s tool kit consisted mainly of what could be characterized as a quantitative genetics-based knowledge, and the education and training of plant breeders revolved around developing this knowledge. New laboratory-based tools. The discovery during the 1950s and 1960s of the structure and roles of DNA and RNA led to the develop- ment of molecular genetics and gave rise to a wealth of new genetic information. By the late 1980s, this information had been used to develop powerful new breeding tools. Inheri- tance could now be linked to the presence of specific nucleotide sequences in successive generations. Variations in these sequences at different physical locations (loci) in the genome of an individual generally give rise to different genes, while variations at a single locus give rise to different forms of the same gene, or alleles. For the first time, breeders were able to relate variation in phenotype to variation at the DNA sequence level. This meant they could contemplate breeding at the sequence level, rather than at the level of the whole plant genotype seen through the somewhat murky lens provided by the plant phenotype. To use and exploit DNA-based informa- tion for crop improvement, breeders must acquire knowledge and skills that relate to the identification and manipulation of specific DNA sequences. Three major areas of activ- ity can be distinguished whose successful use requires command of extensive technical knowledge and proficiency in complex labora- tory techniques. Genomics deals with the physical structure of the genome (structural genomics), as well as with gene products and gene interactions (func- tional genomics), and is built on a foundation of molecular genetics, automated laboratory tools, and bioinformatics. Transformation, the process through which genes can be moved across species boundaries, requires identifying and isolating the desired gene, transferring the gene into the genome of the target plant, usually using a bacterium as a vector, and then regenerating entire fertile plants from the transformed tissue. It forms the basis of the burgeoning adoption of genetically modified crops. Marker-assisted selection (MAS) involves the transfer of a piece of DNA (including the gene of interest) associated with a specific phenotype, with molecular markers being used to identify the DNA fragment in successive