| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Commentary: Growth of beanbag genetics
Genetic Epidemiology, Human Genetics Division, School of Medicine, University of Southampton (Mailpoint 808), SO16 6YD, UK. E-mail: nem{at}soton.ac.uk
Accepted 20 December 2007
Half a century ago Ernst Mayr expressed a zoologist's antipathy to mathematical theories of genetic variation and evolutionary change.1 He later stigmatized them as beanbag genetics,2 for which the relation between genotype and phenotype is fully specified in terms of a small number of parameters. This prompted a spirited defence from JBS Haldane3 who went beyond Mayr's interest in systematics and the origin of species to include beanbag models of population genetics that are increasingly amenable to observation and experimentation. On the contrary, vague concepts like homeostasis and canalization4 do not lend themselves to beanbag models5 and have not survived into the current century. The most striking failure of a vague model was provided by Lysenko, who sacrificed the validity of his theories and experiments to ruthless ambition that led to fabricated data, death or suppression of genetic and agricultural scientists, and decreasing food production.6 Meanwhile, Stalin and many other Marxists professed admiration for the fictitious increase in productivity claimed by Lysenko.
The events described above took place after Watson and Crick demonstrated DNA structure in 1953 but before DNA polymorphisms were proposed for linkage maps in 1979. Their success spawned the Human Genome Project that produced a tentative physical map of our genome in 2001. This led in turn to the HapMap Project that examined single nucleotide polymorphisms (SNPs) in four small samples, now being typed for other markers. The potential to identify genotypes with significant effects on complex diseases is beginning to be exploited in ways that were inconceivable a generation ago. This novelty challenges beanbag genetics as well as other disciplines on which personalized medicine depends. One of these is genetic epidemiology,7 which developed as a reaction against hereditarian bias in population genetics and environmental bias in epidemiology. Confronted with diseases of complex inheritance both genetic and environmental, the parental disciplines were equally at a loss. Unlike evolutionary genetics, where some parameters may not be estimable, the other branches of population genetics and epidemiology require both model specification and parameter estimation. Can multilocus risks meet the requirements? Only if the answer is positive will beanbag genetics continue to provide useful approximations and novel insights. If the answer is negative, we have at present no better alternative. Parsimony rules for the foreseeable future.
Whatever the outcome, genetics and epidemiology may not share success or failure with complex disease, because their approaches have been different. Genetics introduced the tools of segregation, linkage and association studies, which function best when affection has a simple origin. Power to identify causes decreases with their complexity, which geneticists traditionally try to reduce. On the contrary, increased complexity is welcomed by a branch of epidemiology that pursues Developmental Origins of Health and Disease (DOHaD), illustrated by correlation between body size at birth with coronary heart disease and/or type 2 diabetes later in life.8 Within an ethnic group body size at birth is shared more by maternal than paternal relatives,9 but reciprocal racial crosses have similar birth weight suggesting that paternal factors are confounded with maternal diet and health during pregnancy.10 A significant correlation between weight at birth and an appropriate measure of subsequent risk does not distinguish between environmental and genetic causation of a highly heterogeneous disease group and therefore does not encourage studies of families or case/control pairs with whole-genome typing. Until a beanbag model is proposed, its adequacy cannot be tested.
The history of beanbag genetics is encouraging, but its future is as uncertain as for a newborn baby. However, risks determined by a genotype are more easily related to causes and potential avoidance than are adult risks associated with fetal and infant growth patterns. More generally, environmental factors are inaccurately measured and less promising than genetic markers for personalized medicine.11 Current evidence supports simple expression of multilocus risks with little interaction or non-additivity.12 The future will determine whether beanbag genetics will remain the most promising direction for genetic epidemiology.
Conflict of interest: None declared.
 |
References |
|---|
 Top References |
|---|
1 Mayr E. Where are we? Cold Spring Harbor Symp Quant Biol (1959) 24:1–14.
2 Mayr E. Animal Species and Evolution (1963) Cambridge: Harvard University Press.
3 Haldane JBS. A defense of beanbag genetics. Perspect Biol Med (1964) 7:343–59. Reprinted Int J Epidemiol 2008;37:435–42.[Web of Science][Medline]
4 Lerner IM. Genetic Homeostasis (1954) New York: John Wiley and Sons, Inc.
5 Crow JF. The beanbag lives on. Nature (2001) 409:771.[CrossRef][Medline]
6 Soyer VN. Lysenko and the Tragedy of Soviet Science (1994) New Brunswick, New Jersey: Rutgers University Press.
7 Khoury MJ, Beaty TH, Cohen BH. Fundamentals of Genetic Epidemiology (1993) New York: Oxford University Press.
8 Eriksson JG. Epidemiology, genes, and the environment: lessons learned from the Helsinki Birth Cohort Study. J Intern Med (2007) 261:418–25.[CrossRef][Web of Science][Medline]
9 Morton NE. The inheritance of human birth weight. Ann Hum Genet (1955) 20:125–34.[Web of Science][Medline]
10 Morton NE, Chung CS, Mi M-P. Genetics of Interracial Crosses in Hawaii (1967) Basel: S Karger.
11 Mushiroda T, Ohnishi Y, Saito S, et al. Association of VKORCI and CYP209 polymorphisms with Warfarin dose requirements in Japanese patients. J Hum Genet (2006) 51:249–53.[CrossRef][Web of Science][Medline]
12 The Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature (2007) 447:661–78.[CrossRef][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. D. Smith 'Something funny seems to happen': J.B.S. Haldane and our chaotic, complex but understandable world Int. J. Epidemiol., June 1, 2008; 37(3): 423 - 426. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||