And Mike dove in, and we dove in, to collect families on the autism spectrum. Families, we mean one proband, one sibling who's definitely not on the spectrum, and both biological parents who are not on the spectrum. So, you can extract their DNA and compare the copy numbers. And any variant that turned up with too few or too many copies in the proband was immediately suspect. There was some cases where they couldn't be identified with absolute accuracy, but some where there was no question – there was a variant.

The first one that hit our news was the 16p11.2 variant. There was both a deletion and a duplication and they both had certain psychiatric disorders. The deletion was more likely to have autistic-like symptoms or profound anxiety. If not the strict definition of autism, it was somewhere on the autism spectrum. The duplication had a very different phenotype, but it had a phenotype. So, there was a lot of attention devoted to 16p11.2 and by now there are about 300 cases of this rare disorder.

At about the same time, the DSM-5 came out and Cathy Lord, who was a big factor in all of our studies, led the charge to say we should change the diagnosis from autism to autism spectrum disorder, so ASD is the going diagnosis. About the time the genetics became clear, it couldn't be narrowly restricted to the core symptoms of autism. But the movement disorders had to be taken into account, both fine and the coarse movements, beyond the social withdrawal, the profound anxiety and the hyperactivity, how do we take it into account? Language disorders, how do we take it into account? Speech? So, all this fit with the copy number variants, but with time other mutations are turning up. The more people are assayed, that seem to be unique to the autism spectrum. To me, in my own view, the spectrum is getting so broad, the genetics are getting so wide, more thinking has to go into what's going to help us understand the pathogenesis. It's not clear to me and I think that'll be the challenge of informatics and databases. And I think a lot needs to be done on the phenotypes. Sole concentration on genetics isn't going to help solve their 20-year limit of what is this disorder. And it's going to be as broad as all of neuroscience, I think.