The autism gene problem

I wonder why we are still asking questions like "How do you find genes for a disorder that eludes definition?"¹ You even suggest the solution when you quote Steve Scherer saying that "while...neurolignin could be a candidate in some rare cases, other researchers have countered that the family where the gene was found may have mental retardation rather than autism." The answer is to look at candidate genes and find out what symptoms are phenotypically unique to people who have a particular variant of the gene compared with people who don't have that variant.

With the availability of genotyping technology we can now look at groups of people who have, maybe an even more difficult disease to diagnose, schizophrenia, and categorize them into groups based on their genotypes. Then we can look at those groups and, from the similarities of phenotype within the group, and the unique differences...