IntroductionEgenis co-director Steve Hughes's thoughts on 10 years of human genome sequencing.
A pair of recent high-profile accounts of the achievements of human genomics in relation to the public’s health prospects 10 years on from the announcement of the first draft sequence present rather contrasting takes (writes Professor Hughes).
Nicholas Wade writing in the New York Times asks, none too subtly, where are the cures? “Geneticists are almost back to square one” concerning the treatment of common disease, he claims. Hamburg and Collins, writing in more optimistic tone in the New England Journal of Medicine, herald a new network of treatment options just around the corner. All we now need to navigate the world of personalised genomic medicine, they assert, is a bit of translational practice and some regulatory provision.
Wade is persuaded that the assumption that common genetic variants would underpin common susceptibilities, an assumption which has driven genome wide association studies of large cohorts (GWAS) and the foundation of the HapMap (human haplotype mapping) project, has not been born out, and that in practice common variants explain rather little of an individual’s risk. Rather, it seems that an extensive terrain of rare genetic variants are in play and that these fall below the radar of HapMap. These are not the stuff of easy anticipatory diagnosis in populations nor are they likely to point to targets for rational drug design or pharmaceutical intervention. Even where associations between susceptibilities and common variants do show up statistically the genetic loci tend to be uninformative in relation to the cause of the condition itself. (A notable exception lies in the practice of management of some types of diabetes based on the recognition of variants at specific loci.)Again, not much benefit apparent for the sharpshooters of big pharma.
Wade’s pessimism derives substantive rationale from the failure of GWAS to throw up informative loci associated with the risk of cardiovascular disease (see Couzin-Frankel, Science 328 2010, p1220-1221). As Wade emphasises, it appears that family history narratives are more informative predictors of risk. Cancer and behavioural disorders (like autistic spectrum disorders) similarly resist simple genetic tagging and causal explanation via common variant loci.
Hamburg and Collins’ enthusiasm is based on the discovery of “hundreds of genes… contributing to human illness” and “genetic variability in patients’ responses to dozens of treatments.” They are convinced that large investment in genomic science has created a platform for personalised healthcare and genome-based diagnosis, prognosis and individually tailored treatment. While the promise of those hundreds of genes seems to be illusory except in relation to very rare disorders, Hamburg and Collins’ optimism is perhaps more reasonably underpinned by the successes of pharmacogenomic tags, that is the identification of variants at loci associated with the efficiency of enzymic inactivation and excretion of therapeutic compounds and other toxins. Such variants may be used to predict a patient’s response or sensitivity to a particular drug and to design patient-specific dosage regimes. They may also rescue drugs which fail their final trials phase for use in specific genetic sub-groups of the population. But reterritorialisation and rehabilitation of the failed blockbuster is some way short of what the pundits promised 10 years ago as a revolution in health care.
The response of the genomic medicine enthusiasts to the GWAS deficit and the absence of new cures is to draw attention to the plummeting costs of whole genome sequencing and to reposition their expectations there. Of course, even the availability of total DNA sequences for large cohorts coupled to the power of bio-informatics does not guarantee that it will be possible to resolve the complexity of multiple rare variants and to determine an optimised set or pattern of sets for individual or group risk assessment, but we should not write it off yet. The alternative scenario is one in which to benefit from genomic medicine we all need to have total genome sequences. Either way what is worrying is the escalation of resources which the approach demands both for data capture and informatic analysis plus the issues of confidence limits in interpretations and predictive value. The concern is that we are driven further and further into the Pareto-esque* asymmetry in which 80 per cent of genomic medicine, in order to prove its merits, will be chasing the health concerns of 20 per cent of the population, doubtless the wealthiest 20 per cent. Under such circumstances personalised health care would become the antithesis of public health and we would be entitled to feel disappointed. The disappointment may be offset by the growth in our knowledge of genomes and how they work, particularly in the emergence of a more systematic understanding of epigenetics and chemical imprinting of the genome in relation to evolution, development and the influence of the environment. As scientists we have much to celebrate.
If there is some disappointment at present, perhaps we are all to blame for putting cures and restorative interventions on such a pedestal and allowing this hegemony to set the strategy for genomics. We forget that health is largely a matter of consumption and close our minds to the under-nutrition and other non-medical deficits which afflict the health prospects of the majority of the global population (80 per cent). In this light, the real disappointment of the past decade is the relative prominence and resources accorded to genomic medicine over crop genomics. Investment in crop genomics has been small and dominated by corporate institutions, with public investment limited to a few major crops.
Perhaps we can support a vision of the last decade of human genomics having contributed to cheap, high throughput genomics which will allow crop genomics to catch up and to play its rightful role in promoting the deployment of biological diversity in crop improvement and ultimately in supporting public health.
*Pareto was the Italian economist who noticed that the control of 80 per cent of the wealth by 20 per cent of the population is universal across cultures. The principle has been spun out to many other areas such as 80 per cent of purchases are made by 20 per cent of customers, or more generally 80 per cent of effects are produced by 20 per cent of causes, which brings us back to the hunt for genetic variants of significant effect.