Mapping the Autism Spectrum
A highly heterogeneous condition, autism presents a unique challenge to medical geneticists. Heading the charge to understand the genetics of autism, and making it easier to get published, is Dr Steve Scherer.
Awareness of autism as a genetic, or partially genetic, condition has been on the rise over the past decade. According to Steve Scherer, Director of the Center for Applied Genomics at the Hospital for Sick Children (SickKids) in Toronto, he was in the right place at the right time (and speaking to the right people!) to be able to start piecing together some of the first genomic mapping for the condition. Most recently his group succeeded in defining the copy number variations and other genetic features underlying autism, and are now able to use that information to provide much-needed assistance and genetic counselling to affected families.
You can find the full interview with Steve Scherer in the latest issue of Front Line Genomics magazine
FLG: How did you become involved in autism research?
SS: I worked in the early days of mapping chromosomes. Our group in Canada contributed to the mapping of chromosome seven as part of the Genome Project. Back in the 90’s, when I started my own laboratory here at SickKids in Toronto, a very interesting paper came out of Tony Monaco’s group, who was in Oxford at the time. The paper showed the first linkage in autism to chromosome seven, which was the chromosome we were the experts on. Interestingly, and this is the beauty of the story, the same day that paper came out I got a fax from a family in California that had a son that was autistic. He had a chromosome translocation that intersected chromosome seven in the same region the Oxford group had published their linkage to.
I knew what autism was, but I didn’t now a lot about it. From our paediatric hospital, I found out that we were seeing hundreds of kids with the condition. As a result of this serendipitous communication, and being at the right place at the right time, we started the autism programme here. We moved very quickly. We assembled a lot of genomic data from our existing work on chromosome mapping, and my group was already working on new technology, so again it was somewhat of a perfect storm for us.
We were the first in Canada to get the one megabase CGH (comparative genomic hybridization) microarray. Our very first samples were actually samples with autism. When we first described the phenomena of copy number variation, we were running samples from autistic individuals and their parental controls. It was a mixture of foresight, but also building on being in the right place at the right time.
FLG: What makes autism such a challenge to understand?
SS: Autism is exceptionally clinically heterogeneous and, as we now know, it’s exceptionally genetically heterogeneous. People hear the term ‘autism’, or ‘autisms’, or ‘autism spectrum disorders’, and they often think of it as a single condition. Psychiatrists and physicians were trained for fifty years to think of it as a behavioural condition, so people have been thinking about it as a single entity for a long time now. However, there are over a hundred different disorders, with different genetic names like Rett syndrome, or fragile X syndrome, that can have autism as a primary component of their clinical diagnosis. That in itself indicates that autism is heterogeneous. But the name comes down to where the child first showed up. If they are labelled with autism, or if they come through a clinical genetics route their condition may get the name of the microdeletion syndrome.
From the research side, everyone was focusing on the so-called idiopathic autism – the autism that was unexplained genetically. Now it seems that we’ve come full circle and acknowledging that there is a big mix. About 60-80% of what is reported as autism is yet to be defined. The field is now coming to the point where we don’t just focus on the idiopathic autism that has very, very stringent features of the behavioural classification. We’re starting to study anybody who has an autism-like clinical presentation. It’s been amazing. Around ten years ago we didn’t have any of the genes identified that accounted for these idiopathic cases, and now we do. This is because of the microarrays and then the sequencing. The whole field is now completely empowered with the progress from the last decade, and it’s starting to chip away at the heterogeneity. That’s what our sequencing project, the MSSNG Project (with Autism Speaks and Google), is doing. We’re using whole genome sequencing to subgroup the autisms into their genetic contributors.
FLG: Your approach was a little different to what many were doing at the time. What lead to you taking that different kind of approach to what other people were doing?
SS: The beauty of a genomic approach is that it is hypothesis free, and you get a lot of data. To be able to process these large amounts of data you tend to have to use algorithms to smooth it over, and cherry pick the ‘low-hanging fruit’. But when we’re looking at undiscovered phenomena in genetics, often the answers are in the data that is much harder to interpret.
With copy number variation, we were comparing genome sequence assemblies before anyone else because we were one of only a few groups that had access to the public data and the private data through Celera. Craig Venter, Mark Adams, and Richard Mural of Celera were very helpful during this time. We published a paper in Science in 2003 and another in Nature Genetics in 2006 where we reported aberrations in genome sequences at the copy number and structural gene level. Well, if you remember back around 2000 there were a lot of papers being published criticising Celera versus the public draft sequences, and the public draft versus Celera’s work. Well it turns out that many of the differences people were quick to point out, we later reported as being copy number variables. So rather than mistakes, or errors in quality, what researchers were actually getting angry about was natural biological variation.
It’s insight we had based on our decade-long scrutiny of chromosome seven. A lot of people were seeing what we now understand to be copy number variation, but they were throwing the data away thinking it was just noise. We eventually recognised it for what it was, and it was just the philosophy of looking harder at ‘failed’ experiments and explaining the data. It’s something I still teach my students today – if you’re doing cutting edge science most of your experiments should be failing, but you need to interpret what that failing means. Often it’s those ‘failures’ that lead to the big discoveries that no one else has seen because they’ve just thrown that data away