I have been told that recent studies concluded that the gene
responsible for autism is transmitted by the father and not by the
mother. Is this true? Does this mean than children of a woman whose
brother is autistic are not at risk?
If the answer to the first question is no, do we have an idea about
the probability that the children of the sister of an autistic man
develop autism?

Hi Freddo99,

You question is complicated by the fact that there is no one gene for
autism - there are, in fact, at least seven.  Four of these are linked
to genes found on the autosomal chromosomes 3, 7, 13, and 15; the
other three are linked to the X chromosome (making them sex-linked to
the mother).  The article linked by Scriptor (in the comments) also
suggests a link to a gene on chromosome 2.  All this is to say that
the genetic basis for autism is really complicated, and that (given
the X-linkage), there is support for it being passed through the
mother's line.  It's probably impossible to say what the probability
that the nieces/nephews of an autistic man would be - but since autism
"appears to be the most highly genetic of all psychiatric disorders"
(again, from Scriptor's article) there's probably an elevated risk
over the general population.

Remember, too, that with all the different genes that can contribute
to autism, this disease has a wide range of severity.  Often you'll
see the genetics referred to as "autism predisposition" since having
the gene may make it more likely for a child to be autistic, but
doesn't guarantee that.

I know this isn't as clear an answer as you might have wanted.  I'm
including several websites on the genetics of autism that I think
you'll find interesting.

Online Mendelian Inheritance in Man: Autism
http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=209850
This is a VERY technical article about the genetics and inheritance
patterns for autism.

The MedlinePlus Health Topic for autism
(http://www.nlm.nih.gov/medlineplus/autism.html) offers a lot of
information on the disease.  I always go to MedlinePlus for health
information - it is a service of the National Library of Medicine that
offers links to government and NGO pages for diseases as well as two
online drug guides, a good medical encyclopedia, and a dictionary.

National Institute of Child Health & Development: Autism and Genes
http://www.nichd.nih.gov/publications/pubs/autism/factsheets/index.htm

The Autism Society of America: What is Autism?
http://www.autism-society.org/site/PageServer?pagename=autismcauses
This page discusses how genetics and environment can combine to cause
autism.  In general this website is very informative.

National Institute of Neurological Disorders and Stroke: Autism Fact Sheet
http://www.ninds.nih.gov/health_and_medical/pubs/autism.htm#genetics

GeneTests: Autism
http://www.genetests.org/query?mim=209850 (click on the "reviews" box
by "Autistic Disorder")
This is another technical overview of the disease.

It appears that the only known increase in risk is for siblings of a
person with autism (risk is 5% as opposed to 0.1% for the general
population (from the NINDS page); risk for brothers is 7% but risk for
sisters is %1 according to the GeneTests article.  This is because of
the sex-linkage).  I have not found information on the risk to
relatives separated by two steps (such as nieces/nephews to the
uncle).  It would be a lot lower than the risk to the sibling,
particularly if the mother does not show signs of autism.

If you're really worried about it you can see if there's a clinic in
your area that will test for the genes by going to the GeneTests site
linked above and clicking the "Testing" box by "Pervasive
Developmental Disorders".

So, the answers to your questions are: best evidence suggests that
autism is linked to the mother's genetics, if it is sex-linked at all
(you could have no problems with your X chromosomes but still be
autistic).  This means that the children of a woman whose brother is
autistic are still at risk.  There is no evidence of what their
increased risk might be, but it is most likely somewhat less than 5%
(the risk to a sibling of an autistic person).

Please let me know if I can help out any more with this answer!  I
used the UNC-Chapel Hill "Focus on Clinical Genetics & Genomics" guide
(http://www.hsl.unc.edu/guides/focusonclingen.cfm) to get to the
resources listed above.

Librariankt

This recent article on a new autsim-related discovery might be interesting:
http://www.eurekalert.org/pub_releases/2004-04/msh-agf032904.php

Scriptor

This may be of interest to you:

"Researchers funded by the National Institutes of Health have
identified a gene that may predispose people to developing autism. The
gene, known as HOXA1, plays a crucial role in early brain development.
The study was conducted by a research team in NIH's Collaborative
Programs of Excellence in Autism and was published in the December
issue of Teratology.

...after testing the genes of parents and other family members of
people with autism, the researchers found that symptoms of autism were
more common in people who inherited the variant gene from their
mothers than in those who inherited the same gene from their fathers.
Dr. Rodier said that this agrees with several other studies of
patterns of autism in families; inheritance from the mother seems to
play a strong role in who develops the disorder."

http://www.nichd.nih.gov/new/releases/autism.cfm