One great step. The first gene linked to autism.
April 6, 2012 1:39 PM   Subscribe

The etiology of Autism remains a mystery. However, three research teams have for the first time linked a gene to certain forms of autism. This is a great step in the search of what causes this disease.
posted by dov3 (42 comments total) 8 users marked this as a favorite
Fingers crossed. At this point, my greatest wish is that someday, maybe 10 years from now, there's a drug that does for autism what an anti-depressant does for that condition -- not a cure, but a coping device, something that allows you to cope better, day-to-day, hour-to-hour.

Maybe then my son can tell me mundane things, like whether his bed is comfortable, or whether his shoes fit. Maybe I can stop guessing at stuff like that.
posted by Cool Papa Bell at 1:58 PM on April 6, 2012 [34 favorites]

All three studies also found evidence that the risk of de novo mutations increases with parental age. In an analysis of 51 de novo mutations, Dr. Eichler’s group found that glitches were four times more likely to originate in DNA from the male than from the female. The risk is higher in fathers at 35 than at 25 and seems to creep up with age.

So perhaps the broader societal shift toward having children later in life has a hand in the rise of autism diagnoses. Time to start pumping out kids at 22, people! ;-)
posted by Cash4Lead at 1:59 PM on April 6, 2012 [4 favorites]

Ah, this is a subject near to my heart. For those not familiar with the 'tech: Whole-exome sequencing is a fancy new way to sequence 'all' (most) of the protein-coding genes in an individual fairly quickly and cheaply. When you do this in an individual, you find thousands and thousands of rare variants, so the major difficulty lies in sorting out what's important from what's part of 'normal' variation.

Autism is a neurodevelopmental disorder that is defined as a combo of problems with communication, reciprocal social interactions, and a restricted range of interests (peculiar repetitive behaviors, etc). From twin studies we know that there is a strong genetic component to autism, but which of our ~22000 genes are key players has been tough to sort out. Most people working in human genetics think autism is a descriptive term for a whole host of different underlying conditions.

We have historically talked about autism as a multifactorial condition, ie. the sum total of numerous small contributions from 'minor effect' genes plus a big schwack from the environment. Geneticists tend to say that whenever we are unable to find 'major' genes using traditional methods involving studies of large and/or inbred families.

What's key about this study is that they show that some persons with autism actually have new mutations (ie. that arose in an individual egg or sperm cell) in a single gene as the cause of their autism. The special sauce here is that in the clinic, it is fairly straightforward to perform exome sequencing of a child and both (non-autistic) parents (this is called a 'trio' study), and simply look to see what new mutations (often a small handful) can be found in that child (versus the parents). The majority of kids with autism in the past have gone without a specific genetic diagnosis, so the study constitutes proof in principle that you can diagnose some patients based on a single blood test and careful analysis.

This also dents the multifactorial model for autism a bit in that a single gene appears to be the cause in some individuals (rather than a mishmash of multiple genes plus environment). That won't be the case in all cases however.

Out of their 200 variants, I think they picked the sodium channel they mention as a proof of principle illustrating that their method gives plausible results. Nobody is claiming this to be THE gene for autism, which would be waaaay too easy. Ma nature is not nearly that nice to us!
posted by biochemicle at 2:10 PM on April 6, 2012 [27 favorites]

And yeah, Cash4Lead, you would expect older dads to be at higher risk of having autistic kids. I believe there is data somewheres showing this .
posted by biochemicle at 2:12 PM on April 6, 2012 [2 favorites]

It's a small world. The Shendure lab is literally next door to where I am sitting right now. This is a cool study.
posted by grouse at 2:18 PM on April 6, 2012

So... if we have one of those $99 gene splicing data dumps with all our SNP (rs) listings, would it be in there?
posted by crapmatic at 2:24 PM on April 6, 2012

Uh, I don't know of anybody doing this properly on a direct-to-consumer basis at the moment. There are lots of snake-oil vendors (like 23andMe) out there, but if you look at the legal they all have a clause saying 'for entertainment only'. Me, I would spend the $99 on Guiness and wait for it to go mainstream
posted by biochemicle at 2:31 PM on April 6, 2012

Guinness generally goes mainstream through me in about an hour and a half.
posted by gurple at 2:34 PM on April 6, 2012 [8 favorites]

what should a scientist tell someone who is creeped out by the intersection of "disease" and "personality"?

that we are talking about people who are too severely impaired to operate and that the "autism" they're treating is not the colloquial kind?
posted by This, of course, alludes to you at 2:39 PM on April 6, 2012 [1 favorite]

I didn't know Jack Black was a geneticist!
(That's what I thought when I saw the photo.)
posted by demiurge at 2:39 PM on April 6, 2012 [1 favorite]

The gene in question has been linked to autism in the past. This still seems like grossly overblown "forest fires caused by physics" mainstream science reporting. What specifically does this gene do in the body & brain normally vs. abnormally? How does that genetic difference relate to the observed effects of a diagnosed autistic spectrum disorder?

To say that "highly disruptive (nonsense and splice-site) de novo mutations in brain-expressed genes" are associated with abnormal brain function is sort of, duh? They've got no explanatory mechanism, and since there's no accepted definition of autism disorder in the first place, we're left at saying that a small number of people labeled with a category of brain disorder have similar genetic problems

Anyone have a better source than the NYT?

This also dents the multifactorial model for autism a bit in that a single gene appears to be the cause in some individuals

This seems like a significant correlation/causation jump. Doesn't one need to propose some method of causality before claiming causation exists?
posted by crayz at 2:46 PM on April 6, 2012 [1 favorite]

There are lots of snake-oil vendors (like 23andMe) out there, but if you look at the legal they all have a clause saying 'for entertainment only'.

Maybe a derail, but what makes 23andMe a "snake-oil vendor"? My birthday's coming up and my wife wanted to buy me their kit. I thought it sounded sort of cool, if only because there are some interesting questions about our heritage in my family that 23andMe suggests it's suited to answering. Happy to hear about it in MeMail if you have the time.
posted by mph at 2:50 PM on April 6, 2012

Is there some new mutant gene that is possibly causing the increase in autism diagnoses?
posted by pixienat at 3:00 PM on April 6, 2012

Crayz, I agree the NYT article misses the boat. Again, it's the method, not the specific gene, that's getting them published.

Correlation v. causation: Point taken. This is always a concern in human genetics since research ethics boards tend to frown on breaking genes in humans to see the effects. That's why we have to rely on natural experiments (and the occasional animal model).

The main practical use of this will be finding changes in already-known genes in order to diagnose the specific kid in front of you, whose parents want a concrete answer, no matter how abstract, as to why he/she can't talk or interact with them.
posted by biochemicle at 3:02 PM on April 6, 2012

I guess concrete and abstract are opposites but you know what I mean xP
posted by biochemicle at 3:08 PM on April 6, 2012

My wife and I have enjoyed our 23andMe accounts, mph, but part of that is expectations tempered by having only paid $99 (during a DNA discovery anniversary deal). I'm not sure we'd be as gung-ho if it had been a lot more or we were paying one of the monthly ongoing fees they keep trying to hook us into.

I'm not sure how useful you'll find it for heritage questions. You go back a long way before you get concrete relationship percentages and the racial stuff is pretty broadly painted.
posted by phearlez at 3:09 PM on April 6, 2012

The increase in the autism diagnosis is probably mostly linked to the fact that it wasn't really a diagnosis until the 1960s, and awareness of it has been increasing ever since. Children who would previously have been diagnoses with something else, such as mental retardation or childhood schizophrenia. I haven't seen any studies that demonstrate that there has been an increase in actual cases of autism -- just an increase in diagnoses. In fact, a recent British study looked at a large swath of the population and found that a large number of adults who were undiagnosed were probably on autism spectrum, and it was about the same percentage of the population as children with autism.
posted by Bunny Ultramod at 3:13 PM on April 6, 2012 [6 favorites]

In fact, a recent British study looked at a large swath of the population and found that a large number of adults who were undiagnosed were probably on autism spectrum, and it was about the same percentage of the population as children with autism.

Does anyone have a link to this study?
posted by chavenet at 3:48 PM on April 6, 2012 [1 favorite]

Dr. Eichler’s group found that glitches were four times more likely to originate in DNA from the male than from the female. The risk is higher in fathers at 35 than at 25 and seems to creep up with age.

So older mom = Down Syndrome. Older dad = autism. Thanks, nature.
posted by Justinian at 3:53 PM on April 6, 2012 [2 favorites]

So older mom = Down Syndrome. Older dad = autism. Thanks, nature.

It's our fault, really. We had to, you know, go and evolve and shit. We were perfectly fine living in the cave with the diseases and the occasional Arctodus simus attack. But nooooo! We had to go and invent tools and agriculture and stuff. And then suddenly everyone in the tribe was living past 25. Sheesh!
posted by Cool Papa Bell at 3:59 PM on April 6, 2012 [5 favorites]

23andMe's technology won't do anything like this, previous studies based on the SNP calls that 23andME uses weren't successful in the past of finding any genes associated with autism.

If you're interested in exome sequencing, it's becoming affordable. I work in this field too, and get spam from companies for exome sequencing, the cheapest of which was $700 for the first sample, and $1k for bulk orders. But it was low quality, and not nearly enough data for what is currently considered good for variant calling. Best of all the front page of their website in contains large images with only text, in Comic Sans. I kid you not. A font shouldn't be enough to nix a supplier, but when it looks like they're doing a shoddy job, it certainly doesn't help.

In the end, this data is worthless without interpretation, and there's little that your genome sequence can tell you at this point. And the best estimates of geneticists are that your genome is not going to tell you much about your potential or the way you're going to die, or anything of consequence, except for a very very small fraction of people. I did 23andMe, and was thoroughly underwhelmed, and felt like it was a complete waste of money. But being in the field, I felt compelled to at least try it.

Proper full-genome sequencing is about $3k in marginal costs right now, and falling like a rock. I've said it before, but this technology makes the computer tech's advancement look slow. Despite this, it's going to take an awfully long time to pull meaning out of the data, and dissect more and more of the human cellular network and disease. The part of Gattaca that we'll get really quickly is the cheap and easy sequencing, but it's unlikely that we'll ever get the social aspects, and it will be a long time before we get the full health benefits of the technology.
posted by Llama-Lime at 4:02 PM on April 6, 2012 [3 favorites]

This psychiatrist is controversial:

New Studies: Nutritional and Pathologic Clues in Autism

My amateur opinion is that she is on the best track.
posted by bukvich at 4:04 PM on April 6, 2012

As an autistic, i have two thots

a) i don't think there is one gene, and this is not the gene.
b) i really really worry that there will be presurre to push the autism out of me, in a eugenics way.

autism is who i am as a person, and i think of it as a gift
posted by PinkMoose at 4:12 PM on April 6, 2012 [8 favorites]

It is really important to remember that no matter what loathsome groups like Austim Speaks want to claim, autism diagnoses are on the rise, not necessarily autism cases. Up until very, very, very recently -- like, the past half-decade -- the vast majority of people on the autism spectrum would never have received a diagnosis of autism. Many adults on the spectrum still don't have that diagnosis, even people who desperately need help and social services. It is bad science to talk about "the rise in autism" when the goalposts for what autism actually is keep changing so drastically every few years! And they're going to change again very soon when the DSM-V finally comes out, which will put a different emphasis on social function versus other symptoms, which will reshuffle people's diagnoses yet again.

Furthermore, most of these recent studies on the so-called uptick in the number of kids with autism are usually looking at data collected from public schools to see how many kids in a district are using the educational label of autism on their IEP's to receive services, and then extrapolating out from there to a medical diagnosis. But an IEP label used in a school setting does not always match a medical diagnosis. There's a big difference -- often influenced by issues of payment for special education services -- and it's causing dirty data.

My own four-year-old son is one who fits in this category, for example. Two different developmental pediatricians (one at UCLA) confirm that he has multiple developmental delays, particularly in speech and language and fine motor skills, and therefore has communicative problems, but doesn't really fit the diagnosis of autism. But his IEP team through the Los Angeles Unified School District disagrees, so when he goes to kindergarten in a year and a half, he will likely get the educational label "Autism" on his IEP in order to receive services like occupational therapy. And from there he will be counted in these new autism studies, and therefore accidentally be used to justify the ZOMG, MOAR AUTISM meme.

Believe us, we'd actually rather our son be properly labeled with his severe speech and other issues as his primary educational labels, but that's probably not going to happen. This is a complex area that covers issues like how special education is handled by various districts. But frankly, it's partly because if he gets the "Autism" educational label, LAUSD can (and has) try to direct him into their all-autistic-kids special day classes run by their system -- which means they will get much more funding from the state! They tried to shuffle our son into one a year ago, where I had the teacher herself actually tell me that she suspected some of the kids in the district's class are not actually autistic but just language-delayed because they speak two or more languages at home and school and were just behind. But if they get the label, and they're in the class, the district makes money...

This is the second kid in our family living in LAUSD in the past fifteen years to be incorrectly counted as autistic, since his cousin went through this same crap in the 90's! The cousin is now in college and, though shy and a little hard to understand, clearly not autistic. LAUSD also tried to educationally label the cousin as mentally retarded, because that label would also get them more state money...I don't know how they hand-waved away the fact that the cousin then started taking college-level math while still in high school...My point is that this kind of crap has been going on for a long time, the crossover between state money and local money and educational labels and medical diagnoses. It can all be messy and should not be confused for an actual unbiased source of information about medical problems that may or may not be on the rise in your area.

So the incorrect educational label gets applied for whatever reason -- and the kids get counted as autistic in a study -- and the study gets touted and re-blogged without any critical eyes -- and these "number of kids with autism" goes up, oh noes! -- and awful greedy groups like Austim Speaks can continue raking in funding for eugenics research programs designed to prevent autistics from ever being born at all, rather than spending a fucking dollar on actual at-home care or help for actual already-existing autistic people -- and so it goes...

So, to sum up:

- what autism "is" is still, to this day, highly in flux
- older people with autism are rarely counted in any stats as having autism
- younger people with non-ASD developmental problems are wrongly counted in stats as having autism
- educational data is dirty and should not be confused for medical data
- educational labels meant to justify whether a federal, state, or local group has to take responsibility for educational funding often have little relationship to actual medical issues
- eugenics is bad, yo
posted by Asparagirl at 4:15 PM on April 6, 2012 [18 favorites]

Jinx, PinkMoose! :-)
posted by Asparagirl at 4:15 PM on April 6, 2012

More on the study on a NIMH article.
posted by francesca too at 4:21 PM on April 6, 2012

This isn't the study, but discusses it.
posted by Bunny Ultramod at 4:21 PM on April 6, 2012

The study I mentioned.
posted by Bunny Ultramod at 4:23 PM on April 6, 2012

autism is who i am as a person, and i think of it as a gift

And that's great, but this came up the last time there was an autism discussion and you were defensive on this subject and you said this stuff, so I'll repeat what I said: Your autism is not everyone's autism. Your experience is not everyone's experience.

My autistic nephew, who will be five in a week or so, is: not potty trained; does not talk (knows a few words but rarely uses them); makes eye contact maybe once a month; has pretty poor motor control (walking is more like a controlled constantly-falling-forwards) and hand strength; is hyperactive to an incredible degree (will run, nonstop, for 18 hours, pass out for 4, and then do it again, unless medicated into a semi-normal sleeping pattern); stims like mad; and this is the stuff I see when I watch him for 6-8 hours a week, and there're odds and ends of other behaviour that I hear about but don't see regularly (and we haven't even touched dietary nonsense). Some of this may get better with time and treatment, and some of it may not.

His is not the worst case of autism. Not by a long shot.

We know there's perfectly normal stuff going on, locked away in his brain, because he can work an iPad and a computer and a Wii just fine, figure out puzzle games (if they don't involve ANY reading) and the like on his own, and he smiles and laughs a lot (tantrums don't happen too much with him, thankfully). But if someone came out with a pill, injection, magic spell, whatever, to change this stuff? To fix him? In a heartbeat. It is not a gift. It is genetics gone askew.

I'm totally okay with autism being a spectrum, but for every "my autism is who I am" or self-identified Aspergers, please remember there's a kid who's rocking himself in a corner, occasionally arranging blocks in order of size, and his parents and family love him and there is not a damned thing anyone can do for him yet.
posted by curious nu at 4:41 PM on April 6, 2012 [25 favorites]

Totally agree with Asparagirl. The application of a diagnosis has to be first and foremost for the benefit of the child (adult) in question, in terms of getting the right developmental supports etc.

Stupidly, these days, you happen to get waaaaaay better access to developmental services (speech therapy, etc) with a diagnosis of autism than you would with a diagnosis of 'global developmental delay' (formerly 'MR'). Schools get extra funding too, as Asparagirl points out. This means that a lot of nonautistic kids with other sorts of developmental delay do get lumped into a diagnosis of 'autism' for practical reasons. Eventually we'll invent a new term identifying the same group that was originally referred to by 'autism', in which case we will have come around full circle.

I'd be willing to bet the birth incidence of actual Autism has been the same for hundreds of thousands of years. That's really a guess because nobody truly knows though.
posted by biochemicle at 4:44 PM on April 6, 2012

I'm doubtful that the existence of a useful treatment for severe autistics will result in medical professionals pushing such a treatment on those of us who occupy the higher-functioning areas of the autistic spectrum.

I've met enough fellow members on the lower-functioning side to wish there could be something that could help them; our own fMRIs and neurological assessments have helped provide data for autism researchers.

We high-spectrum folks may not desire a 'cure' (and get angry with neurotypical humans for focusing on one), but a treatment that could help severe autistics gain the cognitive ability to decide if they want one would be an overwhelmingly positive step.
posted by malusmoriendumest at 5:21 PM on April 6, 2012 [5 favorites]

I am rereading Dawn Powell's* diaries, and the descriptions of her son's personality, challenges in adapting to daily life, and fascinating pockets of talent seem in key with some of the self-descriptions I've read by autistic people like Daniel Tennant and Kim Peek. Unfortunately, Powell's son spent his life mostly in mental hospitals, because that was pretty much the option for someone whose impulse control issues and challenges with skills of everyday life--despite some strong intellectual gifts in certain areas--were so profound.

I don't know how many other people were born in the late 1920s who had experiences that were like Powell's son (Joseph Gousha, Jr.), but my guess is that there were an awful lot whose lives weren't as well documented by a loving mother who was a professional writer and compulsive diarist. Gousha certainly seems to have had quite a few friends (and enemies; like his mother, he was someone who loved and hated strongly) who he considered to be "like him" at the various institutions where he lived.

*US novelist, 1896-1965, spent her professional career almost exclusively in Manhattan.
posted by Sidhedevil at 5:36 PM on April 6, 2012


based on 'thots' i am pretty sure he is being all ironical like

which is kind of disheartening because while i am not personally afflicted it is nice to see people examining a difficult issue critically
posted by This, of course, alludes to you at 5:38 PM on April 6, 2012

The team found that two unrelated children with autism in the study had de novo mutations in the same gene — and nothing similar in those without a diagnosis.
“That is like throwing a dart at a dart board with 21,000 spots and hitting the same one twice,” Dr. State said. “The chances that this gene is related to autism risk is something like 99.9999 percent.”

This is amazingly wrong. Due to what is commonly called the birthday paradox, the odds of this happening purely by chance are actually greater than 61%.
It's because the odds of all the numbers being different are 21000 factorial over (20800 factorial times 21000^200) which is about 0.3864.
See this webpage for details on the math.

What makes it even more wrong is that they counted multiple de novo mutations per child, which increases the likelihood of collisions even further.
posted by w0mbat at 5:43 PM on April 6, 2012 [1 favorite]

Note that the study had 200 children - you need that number for the math.
posted by w0mbat at 5:44 PM on April 6, 2012 [1 favorite]

I just posted about this in the last autism thread, but it's worth repeating here. If you're interested in these issues, The Thinking Person's Guide to Autism is well worth a look. Contributors include researchers, medical, psychological and educational professionals, parents of autistic children, and adult autistics. Along with many other autism-related topics, the dispute between so-called "high-functioning" adults and parents of so-called "low-functioning" children gets thoroughly hashed out there.
posted by Daily Alice at 6:33 PM on April 6, 2012 [2 favorites]

This, of course, alludes to you

I can't speak for the severely impaired, because I am not; I managed to join the military, pass basic training, and get discharged for nothing remotely related to my ASD, which was undiagnosed at the time.

But speaking for myself, whatever small benefit my introversion and capacity to focus on a single pursuit to the exclusion of all others has given me, I would surrender it in a second for the ability to look someone in the eyes without having to force myself to do so; to speak extemporaneously in full sentences, without frequent, awkward pauses, to go to a bar without needing to stuff earplugs in my ears just to make the sensory assault manageable, to develop more genuine friendships and relationships instead of creepy sycophancies.

I could go on, but I'm not comfortable assigning every social pathology I possess to ASD; it'd be too easy to use it as a crutch rather than attempt to work on them as best I can.
posted by The Confessor at 6:51 PM on April 6, 2012 [1 favorite]

Is there a funding cycle coming up? Is this an example of well intentioned, sincere researchers putting out the best story based on an observation which is most likely to have nothing to do with autism?
posted by epo at 2:19 AM on April 7, 2012

This is amazingly wrong. Due to what is commonly called the birthday paradox, the odds of this happening purely by chance are actually greater than 61%.
w0mbat, as shaky as the leading journals are when it comes to peer review of flashy papers, I think you can probably expect these guys to know the birthday paradox, counting in general, and have a deep understanding of more probability than most hard scienctists. They are geneticists afterall, that's their bread and butter.

The actual number they're using are the nonsense and splice site SNVs, of which there were only 15 in 200 individuals. Nonsense mutations are guaranteed to lop off a large chunk of the protein, and often the entire protein through nonsense mediated decay. Splice site mutations either cut out or add in large chunks of the protein. The other 110 potentially functiona mutations they found in probands were missense, which changes a single amino acid residue in the protein, and may have a huge effect to no effect on the function of a protein.

SCN2A is responsible for gating electrical currents in cells, so prime facie it looks like a highly suspicious gene for any brain signaling related phenotype, and is very likely at least partially a causal gene in these two people. I'm kind of shocked by the unfounded skepticism here. Nobody is saying that this one gene causes all of autism, or even anything more than its an extremely likely candidate for these people.

From these numbers of mutations, it's clear that very little of autism is being explained. Clearly at least some of the explanation lies in the environment, some in random chance during development, and some in molecular changes in proteins (what was thoroughly examined here), some changes in RNA genes (largely unexplored and seemingly very prevalent in neural development), some in the gene regulatory network (genetics almost completely ignored by these papers, because we know so little about it so far), and some in partially heritable epigenetic molecular changes. This looks to be a solid finding that gives far more of a grip on molecular possibilities Han ever before, which is hugely exciting, but it's only a small piece of what will likely turn out to be an extremely complex story.
posted by Llama-Lime at 9:28 AM on April 7, 2012 [2 favorites]

i am pretty sure he is being all ironical like

He isn't. Trust me.
posted by jessamyn at 1:59 PM on April 7, 2012

Lama-lime - in that case the scientist is using a very misleading metaphor to market his work - he explicitly says it's like throwing darts into a dartboard with 21,000 holes, with the clear implication that it's 200 metaphorical darts not 15 (he actually says "a dart", implying one per child in the study, i.e. 200).

Incidentally, with 15 metaphorical darts, the odds are about 1 in 200, still way more likely than was claimed, and a clear indication that the math in this study was not checked by somebody competent.
posted by w0mbat at 2:34 PM on April 7, 2012

They claim odds of about 1 in 125 in this study. You're also not taking into account mutation rate, gene length or GC content, which are all essential for assessing the chances of a given gene being hit twice. In my field, TTN and MUC16 are massively long genes that are extremely frequently mutated in cancer, but due to their large size they are mutated no more than one would expect due to random chance. Sometimes a new grad student will take a simplistic view and estimate significance by simply counting events and get excited about finding a new cancer gene, but by leaving out essential physical details of the system, one has a much less accurate view. In short, the birthday model is not an appropriate or accurate null background, and leads to false conclusions.

I haven't redone their calculations, but the extensive diagnostics that they show in the supplemental information feel correct to an order of magnitude based on my experience, and they're showing all their work. Without a specific flaw in the calculation or the model it would be foolish to doubt that this is a statistically significant result. There's a chance that the scientist was misquoted or misspoke on the 99.9999%, as I don't see a p- or q-value in the paper that corresponds to that, but one can be certain that the scientist wasn't allowed to fact check quotes for this article. The news article is the journalist's creation and is the product of non-specialists using a procedure that is hostile to the accurate reporting of science.
posted by Llama-Lime at 3:30 PM on April 7, 2012

« Older The Real Deal.   |   "One of the most brazen moves in the chess world... Newer »

This thread has been archived and is closed to new comments