“Happy families are all alike.; every unhappy family is unhappy in its own way”. Thus beginneth Anna Karenina. That wasn’t supposed to happen with hereditary disease. The examples we had before large scale DNA sequencing became cheap were basically one gene causing one disease. Two of the best known cases were sickle cell anemia and cystic fibrosis. In the former, a change in a single position (nucleotide) of DNA caused a switch of one amino acid (valine for glutamic) acid at position #6 in beta hemoglobin. In the latter, all mutations have been found in a single gene called CFTR. 85% of known mutations involve the loss of 1 amino acids. But by 2003 over 600 different mutations accounted for only part of the other 15%. There is plenty of room for mutation as CFTR has 1480 amino acids. The kids I took care of in the muscular dystrophy clinic all turned out to have mutations in the genes for proteins found in muscle.
Why not look for the gene causing schizophrenia? It’s a terrible disease (see the post “What is Schizophrenia really Like?” ) with a strong hereditary component. There was an awful era in psychiatry when it was thought to be environmental (e.g. the family was blamed). Deciding what is hereditary and what is environmental can be tricky. TB was thought to be hereditary (for a time) because it also ran in families. So why couldn’t schizophrenia be environmental? Well, if you are an identical twin and the other twin has it, your chances of having schizophrenia are 45%. If you are a fraternal twin your chance of having it are 3 times less (15%). This couldn’t be due to the environment.
It’s time to speak of SNPs (single nucleotide polymorphisms). Our genome has 3.2 gigaBases of DNA. With sequencing being what it is, each position has a standard nucleotide at each position (one of A, T, G, or C). If 5% of the population have one of the other 3 at this position you have a SNP. By 2004 some 7 MILLION SNPs had been found and mapped to the human genome. So to find ‘the gene’ for schizophrenia, just take schizophrenics as a group (there are lots of them — about 1% of the population) look at their SNPs, and see if they have any SNPs in common.
Study after study found suspect SNPs (which can be localized exactly in the genome) for schizophrenia. The area of the genome containing the SNP was then searched for protein coding genes to find the cause of the disease. Unfortunately each study implicated a different bunch of SNPs (in different areas of the genome). A study of 750 schizophrenics and an equal number of controls from North Carolina used 500,000 SNPs. None of the previous candidate genes held up. Not a single one [ Nature vol. 454 pp., 154 – 157 ’08 ]
As of 2009 here are 3,000 diseases showing simple inheritance in which a causative gene hasn’t been found. This is the ‘dark matter’ of the genome. We are sure it exists (because the diseases are hereditary) but we simply can’t see it.
There is presently a large (and expensive) industry called GWAS (Genome Wide Association Studies) which uses SNPs to look for genetic causes of diseases with a known hereditary component. One study on coronary heart disease had 23,000 participants. In 2007 the Wellcome Trust committed 45 million (pounds? dollars?) for studies of 27 diseases in 120,000 people. This is big time science. GWAS studies have found areas of the genome associated with various disorders. However, in all GWAS studies so far, what they’ve picked up explains less than 5% of the heritability. An example is height (not a disease). Its heritability is 80% yet the top 20 candidate genetic variants identified explain only 3% of the variance. People have called for larger and larger samples to improve matters.
What’s going on?
It’s time for you to read “Genetic Heterogeneity in Human Disease” [ Cell vol. 141 pp. 210 – 217 ’10 ]. It may destroy GWAS. Basically, they argue that most SNPs are irrelevant, don’t produce any functional change, and have arisen by random mutation. They are evolutionary chaff if you will. A 12 year followup study of 19,000 women looked at the 101 SNPs found by GWAS as risk variants for cardiovascular disease — not one of them predicted outcome [ J. Am. Med. Assoc. vol. 303 pp. 631 – 637 ’10 ]. The SNPs haven’t been eliminated by natural selection, because they aren’t causing trouble and because the human population has grown exponentially.
There’s a lot more in this article, which is worth reading carefully. It looks like what we’re calling a given disease with a known hereditary component (schizophrenia for example) is the result of a large number of different (and rather rare) mutations. A given SNP study may pick up one or two rare mutations, but they won’t be found in the next. It certainly has been disheartening to follow this literature over the years, in the hopes that the cause of disease X, Y or Z would finally be found, and that we would have a logical point of attack (but see an old post titled “Some Humility is in Order”).
Is there an analogy?
200 years ago (before Pasteur) physicians classified a variety of diseases called fevers. They knew they were somewhat different from each other (quotidian fever, puerperal fever, etc. etc.). But fever was the common denominator and clinically they looked pretty much the same (just as dogs look pretty much the same). Now we know that infectious fever has hundreds of different causes. The Cell article argues that, given what GWAS has turned up so far, this is likely to be the case for many hereditary disorders.
Tolstoy was right.