FINDING OUT WHAT GENES DO
How will sequencing the genome help?
Completion of the awesome task of sequencing the genome has been greeted in some quarters as if it provided the answers to the questions of understanding how genes operate. On its own, it will not do that. It has been described, reasonably, as the book of life. However, it is a book in which we have identified all the words, and how they fit together, but we have little notion as to the meaning of the words. Nevertheless, sequencing the genome will help in a very important way because, for the first time, we know what genes are present. It may be somewhat like looking for needles in a haystack, but at least we know precisely the number of hidden needles, and their form. That is bound to be of significant assistance. In fact, the gains go far beyond that because we have information on the genomes of other species, and on the effects of numerous protein alterations. That is where bioinformatics comes in (Eisenberg, Marcotte, Xenarios, & Yeates, 2000). In essence, bioinformatics is a way of making multiple links among huge data banks of diverse genetic information. This cannot possibly be done through the operation of a human mind alone making relevant conceptual connections (although that is an indispensable part of the process). Rather it requires massive “data crunching” to bring out connections that had not been anticipated. It is a new technology, and one for which the rules are still being devised. Nevertheless, it is a fast evolving technology, which is crucial in gaining momentum. Much the same applies to the use of DNA arrays and microbeads, in which levels of gene expression can be measured for thousands of genes simultaneously (Brenner et al.. 2000; Lockhart & Winzeler. 2000).
How do genes function?
Let us assume that future research has succeeded in identifying a substantial number of validated susceptibility genes. Where does that leave us, and of what use will this be in improving clinical practice? For all the reasons already discussed, the answer has to be “not much”. That is because identification of a genetic effect is. on its own, uninformative on the crucial question of what genes do. Two somewhat separate issues are involved. First, there is the question of the effects of genes on proteins and, through those effects, on the phenotype (see Rutter, 2000b). Animal studies are crucial in that connection. It is only through experimental gene “knock-out” and “insertion” studies in species such as the mouse (that are genetically quite similar to humans) that we can progress (Flint, 1999; Sibilia & Wagner, 1996; Wicker, Todd, & Peterson, 1995). The object of the exercise is to manipulate the genes in one way or another in order to understand the biological effects, with the aim of determining how they influence behaviour. It needs to be appreciated that, for this to succeed, it is not necessarily essential to create an animal model that closely mimics the human psychopathological disorder. That would be difficult for ADHD, and even more difficult for conditions such as autism or schizophrenia. Rather the need, albeit still quite a difficult one, is to have a model that reflects the genetic alterations, and which gives rise to a behavioural picture that recreates essential components of the human condition. It has not proven easy to do this in the field of internal medicine and it will be even more difficult to do so with disorders such as ADHD. Nevertheless, that is what is needed. The immediate way forward is for molecular genetic studies in humans to lead on to transgenic investigations in other species—and this is just the beginning.
Molecular genetics and transgenics in the near future will be supplemented by proteomics—the study of the interplay among proteins (Pandey & Mann, 2000). This requires the very different discipline of protein chemistry, and other elements of so-called functional genomics.
Let us suppose that all of this is successful (and it has not yet been so in a field of medicine), where does that leave us? The main (and huge) unresolved issue will then be gene-environment interplay (meaning both rGE and G x E). In other words, to the extent that genetic effects are dependent upon some combination with environmental risk factors, epidemiological research that brings the two together will be imperative. This means that a full understanding of genetic risk processes will necessitate accurate identification of environmental risk factors. For that to be possible, research designs must be capable of differentiating environmentally mediated risk from genetically mediated risk, and be able to separate environmental effects on the individual from person effects on the environment. Research designs with these attributes are available (see Rutter, Pickles. Murray, & Eaves, in press-a), and they have succeeded in identifying true environmentally mediated risks
(see Rutter. 2000a). Nevertheless, our knowledge on these is much more limited than is sometimes appreciated.
A good deal of evidence is available on risk indicators, but far less on risk mechanisms. That is. data are available on which environmental factors are statistically associated with ADHD (Sandberg, in press), but not on how they operate. For example, to what extent do associations between parental negativity and ADHD reflect genetic mediation (stemming from, say, personality qualities in the parents representing the adult sequelae of their own ADHD), or the effect of disruptive child behaviour on the parents? In addition, insofar as environmental effects are long lasting, what mechanism is involved in maintaining them? What are the effects of environmental risks on the organism? Are there neural effects? Do environmental effects operate through the same brain processes as do genetic effects, or do they function differently? For example, what are the effects of either or both on cognitive processing or mind sets? Do they operate on the risks for the origin of ADHD, or rather on its course?
It will be appreciated that answering these questions will require enormous general population samples (many tens of thousands, and possibly hundreds of thousands) that include environmental high-risk groups, and the use of high-quality discriminating measures of environmental risk. A considerable challenge remains in devising or adopting existing environmental risk measures so that they can be used in samples as large as those envisaged, as well as having the capacity to index environmental risk in a reliable and valid manner.