Access
This article is part of Nature's premium content.
Published online 9 July 2008 | Nature 454, 154-157 (2008) | doi:10.1038/454154a
News Feature
Psychiatric genetics: The brains of the family
Does the difficulty in finding the genes responsible for mental illness reflect the complexity of the genetics or the poor definitions of psychiatric disorders? Alison Abbott reports.
Every family has its foibles, but this one has more than most. The first member came to researchers' attention in 1968 as part of a genetic survey of juvenile delinquents who had been admitted to Scottish detention centres.
To read this story in full you will need to login or make a payment (see right).
Comments
Reader comments are usually moderated after posting. If you find something offensive or inappropriate, you can speed this process by clicking 'Report this comment' (or, if that doesn't work for you, email redesign@nature.com). For more controversial topics, we reserve the right to moderate before comments are published.
An issue that I struggle with is how effective are GWAS in finding causal genes, in which there may be many different mutations in a single gene, each of low frequency, that result in a disease phenotype. For example the human beta globin gene encodes a protein of less than 150 amino acids, yet there have been over 500 coding region mutations described for this gene, many of which result in a hemoglobinopathy. Another 500 or so mutations have been described in the untranslated and control regions of the gene. The first amino acid in the beta globin protein has 6 different mutations described for it alone. I understand how linkage and family based studies can deal with this problem, but I think (and I'm not a statistical geneticist) that a case control association study would rapidly lose power in a situation like this (many rare mutations in a single gene that in aggregate account for the disease prevalence in the population). If this situation is coupled with a less than ideal phenotype the odds of success seem even less likely.
I think that the genes involved in sporadic mental illness are very varied because there are so many genes that control myelin formation and brain development which are susceptible to mutation due to increasing paternal age. Therefore autism, and schizophrenia are caused by hetergenous mutations. The important factor is to father babies in ones 20s and very early 30s to prevent mental illnesses and mental retardation as well as non-familial Alzheimer's, and some cancers, and autoimmune disorders. Genetic clock ticks for men geneticist Les Sheffield June 12, 2008 12:00am "MOST men would have been surprised to read that overseas researchers had found the death rate of young adults was higher if they had been born to older fathers. This is no surprise to me. It has been scientifically established that genetic changes occur more often in the sperm of older fathers than younger fathers. As men age there is a higher chance of changes in the genes in the sperm. These changes can cause genetic conditions in their offspring, such as birth defects, autism and schizophrenia."
I think schizophrenia is not much more complicated than dysleixa as far as phenotyping is concerned (though I am not a psychiatrist), and I do not understand why the candidate genes for dyslexia appears to be more established than those for schizophrenia. Can someone try to compare these two camps of neuroscientists/geneticists and find out whether and/or why they differ?
Gary has an important point "many rare mutations in a single gene that in aggregate account for the disease prevalence in the population". I would like to add here that deep sequencing of a large number of case and control samples can uncover these rare mutations. The real question however is which gene to sequence! There are two choices here. One, select a gene based on its biological candidacy. Second, pick up a gene based on its statistical candidacy. What constitutes the biological candidacy however challenges the first choice. As regards to second, prior association of a common variant, single nucleotide polymorphism, for example, may indicate the suitability of the gene for deep sequencing. This selection criterion will however still need an explanation as to how a common variation can be found to be associated in the very first place if the real causal variations are going to be rare. No doubt genome level deep sequencing will be most unbiased. Its prohibitive cost will be challenging though.
Alison Abbot is to be congratulated on this most interesting article. I strongly agree with Steven Weinberg that in searching for genes associated with mental illnesses, the gene should be uncoupled with the diagnosis. What I disagree with him is the contention that well defined symptoms like hallucinations or delusions are consequences "far downstream" of faulty genes. These must be as a result of a specific way that the brain malfunctions and therefore be strongly genetically affected. Hallucinations for example maybe experienced by those with schizophrenia, unipolar and bipolar disorders. If one screened separately for genetic associations for these three supposedly distinct illnesses, (as is being done now), one could miss a statistical relationship between a gene and the diagnosis, whereas one could show up if one screened for persons who have hallucinatory episodes or delusions. Maybe his definitions of 'intermediate phenotypes' need to be further examined and refined. I also strongly agree with him that we should screen for candidate genes on the basis of biological reasons for their associations and I strongly disagree with Hyman.S. Mitchell that "We are just too ignorant of the underlying neurobiology to make guesses about candidate genes." If anything we are "just too ignorant" of the underlying neurobiology to make guesses about specific mental diagnoses. When we make a hypothesis in science it is generally not on the basis of a stab in the dark but on intelligent guesses based on the limited experience that we do have. We seek co-relations from our observations and one co-relation that I am pretty convinced about is that mental illnesses run in families and thus there must be genetic links. The abandonment of conventional clinical definitions is a tough call however, and would be strongly resisted by the well entrenched, and largely well meaning psychiatric community, who would then be forced to admit that the science they have devoted their life to studying and using in their practices is substantially wrong. I cant see them lining up for this admission.
The answer to the question ‘does the difficulty in finding genes responsible for mental illness reflect the complexity of the genetics or the poor definitions of psychiatric genetics?’ must surely be both. Any further argument is specious, except that the stakes are now high, with the recent gift of $100million by the Stanley Foundation to the BROAD being more than matched by that of the Leiber Foundation to Danny Weinberger. In the hot air dispute of who is conceptually and strategically right and who is wrong, only time and evidence will tell. But let us not miss a most important point. In a field long starved of substantial charitable or bequest funding, these injections of faith and capital are welcome indeed. Let us hope that others will follow. I for one have no shortage of ideas for how to spend $100million profitably. Not surprisingly, it would start with our discovery of DISC1, but not end there. Rather, we are moving now to integrate the unequivocal genetic evidence from the Scottish family into the DISC1 pathway genetics and biology, in turn leading us directly to neurological systems and processes that make eminent sense in our search for the underpinning of mental disorder and of well being.
It wopuld be pretty good to see if there is any functional correlation between certain genes and whether particular genes work in precise functional modules, ie.e which comibinations are switched on in aprticular individuals. There is redundancy but may these phenotypes focus more on subtlety and quality, as opposed to quantifiable effects pertaining to protein amounts. And then there is the environmental component, which must be taken into account for it's "not all in our genes", as cliche as that might sound. Nonetheless, genes to play a role...an interestying speculation would be whether different environemntal conditions can effect the expression of genes....i.e. the genes may effect the manisfestation of the diease, rather than its inital relaization...not sure tho...it's far too complicated...
to me, I quite like the roll -on, chain - reaction model, a little like the cancer model...that may lead us somewhere. It does seem that the difficulty with findiong genes for various cogntion - related deseases is that they are emergent manifestation of the operation of an orchestra of gens, for it is clear form the amount of fucntional overlap, there is a sort of one gene solution. May be it would be a good idea to look into possible models of redundancy, so we sort of distingusih high - priority gene from low-priority genes and group them together accordingly..
Abbott's thoughtful feature considers how to identify SNP's associated with increased risk of schizophrenia and other psychiatric disorders, despite the failure of previously reported associations to be reproduced in large high-powered studies not limited to a single family. While this discussion is thoughtful, it ignores the possible inadequacy of the simplistic model that individual (or multiple) chromosomal gene mutations are the basis of the known familial incidence of schizophrenia and other genetic diseases. When repeated experiments to test this model have failed, at least some effort should be applied to testing alternative models, rather than only improving or refining the previous failed approaches. One alternative model proposed by Amar Klar (Genetics 162:1259, 2007) is that a specific locus on chromosome 11 determines brain lateralization, by controlling the somatic DNA strand-specific imprinting and selective chromatid segregation of epigenetically differentiated sister chromatids at a very early stage of embryonic development. Normal function of this gene results in lateralization leading to right handedness. According to this model, non-functional alleles of this gene result in random development of handedness, and also results in altered brain development increasing the risk of schizophrenia. Klar has demonstrated that this model explains the inheritance of right and left handedness in humans; furthermore it is consistent with the incomplete but significant association of left handedness with schizophrenia risk. This model is attractive in being consistent with the data and not addressed by the methods previously used. Of course, it is not proven, but some effort should be addressed to testing it, which might be much less expensive than the massive genomic screens repeatedly proposed. Other models consistent with failure of chromosomal SNP analysis to find strong associations with schizophrenia risk could include mitochondrial mutations and heritable prions. When studies based on existing models failed, new models should at least be considered.