Wednesday, June 22, 2005

Genetics 101: "The Amish Anomaly"

Dan Olmsted, senior editor at UPI, has found the cause of autism by looking for it among the Amish people of Lancaster County, Pennsylvania. He claims that, because he couldn't find enough autistic people among the Amish, autism must be due to something that the general population does that the Amish do not. Skipping over SUV's, computers, cellphones, jet airliners and television as possible causes, Mr. Olmsted instinctively homed in on the one "true" cause: vaccines!

Expecting to find 1 in 166 Amish people to be autistic (probably from US Department of Education data, but no source cited), he could only find one. This child, a girl adopted from China, had been (gasp!) vaccinated. This, apparently, was enough of a smoking gun for Mr. Olmsted to indict the mercury in vaccines as the cause of autism (see also), since the Amish generally don't vaccinate their children.

The obvious first question is how accurate Mr. Olmsted's "epidemiological survey" might have been. The Amish are not known to be an extroverted group who welcome intrusive investigative reporters into their homes and tell them the intimate details of their families. It is entirely possible that Mr. Olmsted may have missed a few autistic people. But leaving aside the question of accuracy, is it possible that this "anomaly" of apparent low autism rates among the Amish might have another explanation?

There is another equally plausible (some might even say more plausible) explanation - one that Mr. Olmsted has apparently dismissed, since he had it right in his hands. The Amish, and their neighbors the Mennonites, have been studied by geneticists for some time because they are a genetically isolated community. Although they accept converts, they don't get very many and so they don't get much "new blood" (genes). In addition, they don't move around much and their members tend to marry within the community - those who don't often leave and join their "English" (as they call people outside their communities, regardless of ethnicity) spouse outside of the community.

In genetic terms, what is happening is called "inbreeding" - and it has some very predictable results. One of these is that some variations of genes are lost and other variations become more common. There is a tendency to genetic uniformity, with only a few of the possible variations remaining. Among the Amish and Mennonites, the result has been a dramatic rise in genetic disorders that are more rare in the general population (see here, here and here) as a result of this rise in frequency of the "abnormal" variations of several genes.

Although inbreeding has a generally bad reputation (think Romanov's) because of the rise in genetic disorders, it can also lead (with equal likelihood) to the loss of genes for "bad" things, like autism. So, it seems only fair that, with their increased prevalence of so many genetic disorders, the Amish might have gotten a bit of luck and lost the gene for autism. Of course, this sort of thing doesn't make good UPI articles, and so apparently got flushed with the rest of the rational science that Mr. Olmsted managed to dig up.

Now, I have made this argument before and gotten the response, "Do you really think that the Amish have different genes than we do?" No, I don't, but that question exposes the ignorance of the speaker. All humans, with a few exceptions due to chromosomal deletions or duplications, have the same genes. It is the different variations of these genes - called alleles - that make us all unique individuals (except, of course, for indentical twins, who have the same alleles - sorry, folks).

It is quite reasonable to assume that the Amish have different alleles (more precisely, different allele frequencies) than the general population because they are pretty much genetically isolated from the rest of us. The genetic work done on the Amish and Mennonites confirms this - the only stretch is assuming that this difference will have no effect on the prevalence of a certain disorder, namely autism.



Anonymous Anonymous said...

Has it been proven that there is a "gene" for autism? My understanding was that they are looking at several genes that could be connected to the dissorder but have never pinpointed "one" gene. Current research seems to be along the lines that autism is cuased by a combination of factors ... a genetic disposition combined with some kind of environmental factor(s) Its simply not possible for there to be an explosion of a genetic nature of this magnitude in society. Humans do not reproduice fast enough to see the rises in autism rates that we have seen in the last decade or two and it be attributed purely to genetics. There must be something else.

08 April, 2007 19:40  
Blogger Prometheus said...


Wow! You're really reaching back into the archives - this post is almost two years old. That's like two centuries in 'blog years!

Anyway, there are a few alternative possibilities you may have overlooked:

[1] Autism rates may actually be stable, with the diagnosis of autism becoming more frequent. There are a lot of possibilities for this, including (but not limited to) broadening of the diagnostic criteria (already well-documented) and greater awareness.

[2] If autism rates are increasing, then we need to look at possible triggers. I agree that there is almost no chance that classic genetics can explain a significant rise in autism prevalence in such a short time period (but remember [1]). If it is a genetically-mediated response to an environmental trigger, then we need to look at things that:

[a] have increased dramatically since 1985 and

[b] have remained on the increase.

I'm completely behind looking at environmental triggers, but the ones that people are fussing about don't meet both criteria [a] and [b]. Thimerosal exposure, for instance, was pretty constant through the late 1980's and has fallen off dramatically since 2000, but autism prevalence has continued to rise.

So, if someone comes up with a likely candidate, I'm all for looking into it. Unfortunately, all the effort seems to be going into "proving" a cause that looks very unlikely.


09 April, 2007 14:16  

<< Home