Thanks to all on this thread for some very interesting and sensible
cautionary posts, I agree it is nice to be able to discuss some real issues
here. On the genetic findings, we have indeed been treated to claims about
this and that gene before and things rarely turn out to be so simple. Always
best to wait for the technical comments and followup studies.


On 2/19/07, Ross S. Feldberg <[log in to unmask]> wrote:
> Laura
> Thanks for posting this and perhaps it is an opportunity to get the
> discussion away from conspiracies and AIDS.
> I haven't had time to take a close look at the Nature Genetics paper,
> and it would certainly  be wonderful if the report was true and
> significant, but after more than 30 years of behavioral genetics
> raising false hopes I would point out a few cautions before one accepts
> this work quite yet.
> First, in order to enhance their chance of finding a genetic cause, the
>   investigators had to limit their cases to those in which families
> showed multiple incidences of autism. This is a perfectly reasonable
> approach, BUT we need to know what percentage of the cases of autism
> fit into this model. If only 5% of the families that have an autistic
> child have multiple autistic children you may be studying something
> that, while interesting, may only apply to a minority of the cases.
> Second - there is always the danger of using a single term to describe
> a multiplicity of illnesses. Is autism really only one condition? I
> have no idea, but one needs to be cautious in deciding if any
> information is generally applicable.
> Third - the news release mentions "interaction between several genes"
> and so far it appears that many behavioral problems stem not from the
> strong effect of a single gene (example - sickle cell anemia) but
> rather from the weak effects of multiple alleles plus pre and/or
> postnatal developmental conditions.
> I could continue with several other caveats - and while I hope the
> neurexin 1 really is a key player and that it leads to earlier
> diagnoses and  perhaps even treatments, if history is any guide these
> types of announcements are too often driven by the expiration of one
> grant and the need to convince an agency to fund a renewal. I would
> love to be wrong.
> Ross Feldberg


Michael Balter
Contributing Correspondent, Science
[log in to unmask]