By Grant Jacobs 05/03/2019

Researchers are pounding the trail of the genetics of autism. There is also more evidence that vaccines don’t cause autism. More interesting is what does cause autism. (Those looking for the vaccine bit should track down to the section, Another nail on the coffin.)

Not the Rain Man

People with autism or autism spectrum disorder find social communication and interaction difficult. Austism can also show up in repetitive behaviour such as stacking the cans, ordering toys in rows, or strong interests on particular narrow topics.

Autism occurs in around 1-2% of people and affects 4 to 5 more males than females. It has a broad definition and is often found with other neurological conditions.

Hollywood movies like Rain Man and TV dramas popularised autism. These often idolise autistic savants, not ‘ordinary’ autistic people. Curiously Kim Peek, the real-life inspiration for the Rain Man character played by Dustin Hoffman, was not autistic. Brain scans showed that he had corpus callosum agenesis, where the two hemispheres of the brain don’t communicate.

Earlier forays

Comparing twins can expose genetics. Identical twins share all their genes. Fraternal twins share only a half of them. Comparing them you can get some idea of if a trait is due to shared genetics. You might think twin studies would resolve things, but in the case of autism it doesn’t on it’s own.

Twin studies say something about what genetics passed down, but less about genetic changes that happen after fertilisation of the egg. It also says something about the environment they grew up in.

Twin studies from over 50 years show autism has a strong genetic component.[1] 70-90% of the time one identical twin has has autism, the other does too. It’s also more common in siblings.

Very rare conditions typically have one or a few genetic locations that when altered cause a condition. If they’re caused by changes in only one gene, they’re called single-gene disorders or a Mendelian disorder. Cystic fibrosis is an example. Other examples include sickle cell disease, Fragile X syndrome, muscular dystrophy or Huntington disease.

It quickly became clear autism didn’t have one gene causing it but many. Conditions with many genes causing them are complex genetic disorders (or multifactorial disorders). They’re referred to as polygenic, involving many genes.

They’re much harder to study as their inheritance is simply confusing!

One problem with autism is that it is so varied and associated with many other conditions. Scientists looking for genetic associations need groups of people who are very similar and are likely to have the same condition. It’s hard to do for autism. It’s also hard to get the sheer number of patients needed to find the genetic causes because of this.

Speculative thoughts

The absence of a firm answer leads to speculation. Biologists speculate too! That’s where hypotheses—ideas to test—come from.

While it’s clear autism is mostly genetic, it’s hard to find what that genetics is.

But why is it hard to find the genetics? It could be simply that autism involves many genes. It might also be that less common genetics needs to be considered. Or a mixture of both.

Not only are many genes linked to autism, but their effect might be through them interacting with each other, rather than each acting on their own. Particular combinations of variations of genes (allele or mutations) might cause autism.

Mutations change our genes. De novo mutations arise anew, either in the germ cells (sperm, egg) or the body of the child. They’re not in the parents’ bodies. If we were to sequence the genomes of the parents’ body cells (somatic cells) we wouldn’t find them. The child is the first generation to have them.

A growing embryo copies mutations in germ cells into all cells in the embryo. Somatic mutations only occur in some cells. Then the body is partly mosaic with a mixture of variations of that particular gene.

Copy-number variations might be another feature. They’re where a repetitive part of a gene varies in numbers of copies, or are duplications or deletions of a whole gene or region. Some variations not affect a gene itself, but regions controlling of the use of a gene.

Developmental variation is yet another thing to consider. Our bodies can develop differently. Sometimes this is by essentially sheer dumb luck: something messed up inside us while we were growing.[2]

There are a lot of different things that could be going on. Genetics has a lot of subtle twists; these are just examples. What are researchers finding?

Recent work

Ten years or so ago studies on autism were occasional. Now work on autism genetics is pouring out.

‘Back then’ genetic locations found could explain about 10% of autism. Today genetic locations can explain about 25-35%. It’s slow progress, but it’s progress.

There are now hundreds of possible genetic locations to consider, but the contribution of each is small. It’s as if autism genetics is just out of reach.

Over the last 3 months there are over a dozen papers are worth bringing to non-scientists. Originally this article was to give a overview of these to give a feel for the effort involved.[3] As this article is already delayed coverage of recent genetics—if it will be tackled— is deferred until later.

One paper from January in particular is intriguing, suggesting it is the processing of RNA copies of genes that may play an important role. When our genes are read to be used, they’re read into RNA copies. Those copies are transported to where the proteins our genes encode are made. These RNAs can be edited; they’re suggesting it’s this editing that matters, and that the editing explains the link between some autism cases and Fragile X syndrome.

Maybe that’s the missing key? Maybe the RNA editing will turn out to be complicated by somatic variation? It’s a complex ride, but clues are emerging.

Possible environmental factors

This blog focuses is on genetics. There are many (many!) ideas of possible environmental factors. Some look more speculative than evidenced, others potentially interesting. There isn’t time or space to cover these here.

It’s worth pointing out that some (not all) of these ultimately expose or develop genetic risks. For example, older fatherhood, a known risk factor for autism, increases the chances of de novo mutations in sperm. Age might be technically ‘environmental’, but it comes back to genetics. Environmental and genetic elements can be looking the same thing from different angles.

Similarly, cytotoxic substances—such as from smoking—cause mutations to genes. They come from the environment, but their effect is on the genes rather than metabolism or physiology.

There is evidence that neuronal (brain) differences in autistic people start in utero. Naïvely, in the absence of other factors like smoking when pregnant, if a condition like autism starts very early, it’s more likely to be genetic.[4] (It’s also before any childhood vaccine.)

This isn’t to say that there are no environmental factors for autism, but that care is needed when thinking about them.

Historically one environmental factor has long been ruled out: the idea that unsympathetic mothering (‘refrigerator mums’) causes autism. Another ruled out is vaccines.

Another nail on the coffin: vaccines don’t cause autism

Just out today is another study confirming that the MMR vaccine doesn’t cause autism. It’s the biggest yet, examining over 650,000 Danish children born between 1991 and 1998 (over 6,500 were diagnosed with autism during followup), with a decade of followup. They conclude,

We found no support for the hypothesis of increased risk for autism after MMR vaccination in a nationwide unselected population of Danish children; no support for the hypothesis of MMR vaccination triggering autism in susceptible subgroups characterized by environmental and familial risk factors [for autism]; and no support for a clustering of autism cases in specific time periods after MMR vaccination.

Basically: nada, nil, zip.

There are many earlier epidemiological studies[5] comparing people with and without the vaccine.

In a scientific case of dotting i’s and crossing t’s, epidemiologists have pointed out potential ‘gotchas’ in previous studies that might allow for a small fraction of autism cases to be caused by vaccines. Examining these they find that these don’t contribute to autism either. There really is nothing to it: the MMR vaccine doesn’t cause autism.

Vaccines were never likely to cause autism. These studies are mostly for non-scientists to help make that clear using simple studies that essentially just count people and see what happened to those with the vaccine, and those without.

Alison Campbell over at bioBlog and Helen Petousis-Harris at Diplomatic Immunity have short takes on the recent study. (If there is interest, I could present a breakdown of the paper later).[6] There’s also an editorial on the paper.

The main risk factors for autism

Hviid’s team used a number of (non-genetic) risk factors for autism to check if the MMR vaccine was associated with these in some way. It wasn’t. They checked for the risk factors for autism in their study. The main factors were having siblings with autism (hazard risk, 7.32) and being a boy (HR, 4.02). These two have been known for a long time. Lesser risk factors are being born in a late birth cohort (2008-2010; HR, 1.34) and having no early childhood vaccinations (HR, 1.17). The last risk is modest but might interest those objecting to vaccines. (This may be particularly true for girls where the MMR vaccine showed reduced risk of autism.)

It’s worth noting their study may call a slightly higher risk of vaccines causing autism than is actually present. They use the age of diagnosis of the child not the age of autism onset. In cases a child may have been diagnosed with autism after vaccination, but their autism may have started before their vaccination.

Also worth noting is that these studies leave open that autism might start late. Biological studies suggest autism starts in utero – before any of the vaccines.

A quick word about genetic tests. Be careful with claims that autism can be diagnosed with a genetic test. Many of these genetic variations are not diagnostic for any one person. They show up as more frequent in a group of people, suggesting that variation may have a role in the big scheme of things. Each one variation on it’s own may not say much, even nothing at all.

We can eliminate measles entirely

The vaccine associated with unfound fears of causing autism is the measles, mumps and rubella vaccine. Measles is one of small number of infections that we can potentially eliminate entirely. It’s part of a global elimination program that partly works off the backs of the polio elimination program.

In an upcoming post I’d like to introduce a simple, if contentious, idea that might help.

Other articles on Code for life

Autism is mostly genetic again

A gene drive in mice – but only for females

Genome-edited babies – what’s the worry?

Autism genetics, how do you copy?

Autistic children and blood mercury levels

A summary of the evidence for the main vaccine concerns

For new parents or parents-to-be facing vaccine opinions


  1. Long before Wakefield’s infamous claims.
  2. This variation is considered “environmental” in calculations of hereditary. It’s not environmental in the way that usually means, but how our genetics is being played out. This variation is epigenetic, in the original Waddington sense of the term, not the much hyped epigenetic inheritance, let alone the truly strange “epigenetics” from some people who tout unscientific “remedies” for any number of things (or Deepak Chopra’s idea of epigenetics, either!)
  3. It’s also why this post is late, something I’m gutted about. I worked into the wee hours on this and it’s not as if anyone pays me. (I’d set up something like a Patreon account to help fund my writing work, but realistically would anyone support it?)
  4. Changes can start late and be genetic, too. Many ‘classic’ conditions don’t start until a gene is used later in life. One class of these includes the apparently spontaneous deaths of older children and athletes. In cases these are incorrectly attributed to vaccines by parents or those opposing vaccines. These are a story in their own right.
  5. I’m not a fan of lists of papers—you’re better to read and understand a decent summary of the evidence. But as an example of just how many papers evidence a lack of link between vaccines and autism there’s this list of ~140 papers. It’d take you several weeks to read that lot.
  6. You may get the gist reading it themselves if you understand what hazard risks and confidence intervals are. If not, you’ll get confused! It’s not a long paper.

About the featured image

Human chromosomes, coloured by UCSC genome browser default colours. (UCSC = University of California, Santa Cruz). The genome browser is a tool that lets researchers browse and compare genomes, their genes and their properties.

Source: Wikipedia. Public domain image. Author credit: HYanWong.