Publication of the structure of DNA by Watson, Crick, Wilkins and Franklin in Nature on April 25 1953 was, undoubtedly, a turning point in biology. It revealed the secret of inherited variation and mutation and therefore the molecular basis of Darwinian evolution and Mendelian disorders. However, whilst DNA (and RNA) underpins all life on earth, it is not, in itself, 'the secret of life'. This was an understandable hyperbole in the excitement of the moment, but not 50 years on, please! With the human DNA sequence now 'finished' and released into the public domain (almost unannounced, such was the noise of the 50th anniversary celebrations), it is worth remembering that DNA sequences are the notes of the piano, not the music. We need the same combination of imaginative and meticulous science shown by those DNA pioneers fifty years ago to meet the challenge of the next fifty - the complex, dynamic interplay between genetic variation, gene activity and environmental pressures over the generations.
Clearly, future discoveries will build on past achievements in genetics, but predicting the rate of future progress in medical genetics on past successes is a risky business. Yes, we may have discovered the mutant genes that cause most of the simply inherited (Mendelian) diseases over the last 20 years, and developed reliable tests as part of services for families at risk, but the future is not just more of the same. Unfortunately, discovering the genes/mutations in Mendelian disorders, such as cystic fibrosis and haemophilia, is not reasonable grounds for expecting similar success in complex traits, such as diabetes and asthma.
The Mendelian disorders are, by definition, single locus (monogenic) defects of the DNA because the disease tracks reliably with the segregation of chromosomes within families. If the pattern of inheritance is not so, we don't call it a Mendelian disorder. So by selecting Mendelian disorders we are also selecting disorders where a single causative mutation will definitely be found provided we can analyse all DNA sequences. It is no more impressive than finding the broken-down car blocking a street that is causing traffic chaos in a city, given the means to examine each street systematically. How different is the task of understanding the determinants of other types of traffic jam! What is the knock-on effect of what? Against a long-term trend towards vehicle saturation of the existing roads, there is today's cup football match, the student protest rally, the fact that school is on half term, congestion charges and the mayor fiddling with the traffic-light timings, not to mention the drivers' reaction time, adoption of alternative routes and road rage! Studies are needed that can look at multiple interacting influences over time. Comparable studies of complex, common human diseases and traits represent a huge undertaking, but they are possible where there is commitment and political will.
Even if we do extend biomedical research to include longitudinal, population studies that also incorporate measures of the physical, nutritional and psychosocial environment, are we going about the genetic analysis in the right way? This depends, of course, on the aims of the research. There are fundamental differences between the genetic analysis strategies needed for evolutionary and human population research and those best suited for medical research. In the hype over the latest high throughput genotyping technique, this point is often overlooked.
Much of the hype surrounding the international SNP consortium (documentation of millions of Single Nucleotide Polymorphisms (SNP) - a particular type of common differences in DNA sequence between people) centres on the claim this will give us the power to uncover the genetic influences in common diseases. However, SNPs are relatively stable and ancient genetic variants, the stuff of population differences and evolution, not mal-adaptation and disease. Much genetic influence in disease is like 'a spanner in the works' or 'several spanners in the works'. Medical geneticists are interested in the 'spanners', whilst biologists are interested primarily in the 'works'.
Genotyping strategies designed to clarify the 'works' might be poor at revealing 'spanners'. Sidney Brenner (2002 Nobel Prize winner) says most repeat DNA sequences are 'junk' and don't do anything important ('junk' is rubbish that is kept, unlike garbage that is rubbish that needs to be chucked out). He is likely to argue that there is no point in including such repeat sequences in genotyping strategies, particularly as they are difficult to analyse. However, repeat sequences of the wrong size or in the wrong place can screw up the works. Indeed such variants and relatively unstable DNA sequence mutations underlie a number of disorders. The much trumpeted SNPs (selected in part for ease in automated whole genome analysis) are best suited for 'out of Africa' population studies, but are not ideal for discovering disease susceptibility variants where many are likely to be relatively recent in origin and 'unstable'. The genome industry tends to use medical breakthroughs as the justification for raising and spending the money, but I question whether the hype is justified on most current research strategies.
To be fair, one area where I do think great benefit can be expected is in the study of genetic variants in drug metabolising enzymes (pharmacogenetic variants), because here the environmental exposure (the drug ingestion) is known and relatively simple and the genotyping strategy is supplemented by good functional studies and is not just dependent on genome-wide SNP analysis.
My greatest concern of all is that the justifiable air of confidence that accompanied the Human Genome Project has been uncritically adopted for genetic research into common disease. There is a tendency, always present, to fall into the trap of thinking that the main elements of understanding life or health have been discovered and that all that is now left is filling in the detail. We feel we have an adequate explanation, partly because people with vested interests from teachers to venture capitalists keep telling us so. Why should we expect some new discovery to 'turn everything on its head'? We are back to 'the secret of life' claim, in which the DNA sequence is somehow extracted from the living process to be 'put in charge', where conception is reduced to the mixing of two sets of genes. It was necessary to understand DNA to know the nature of life, but that doesn't mean it is sufficient. The big impacts on the genetics of common disease in the next 20 years are likely to involve processes that are currently unknown or barely hinted at in the daily avalanche of biomedical research data.
As Matt Ridley's book 'Nature via Nurture' illustrates so well, '[genes] are themselves exquisite mechanisms for translating experience into action'. Genes are largely mediators in the life process and for most traits and common disorders genetic influences will be conditional on the environment. And it is beginning to look as if it is not just the individual's lifetime experience and exposures that play a part. There is now good evidence that human sperm carry information about the ancestral environment and that this influences development and the longevity of the descendants. A recent Swedish study indicates that sharing an increased risk of diabetes with your siblings is about the nutrition of your paternal grandfather as well as shared genes. It is time we put genes in their place.
Marcus Pembrey is Professor of Clinical Genetics at the Institute of Child Health, London