> To all the experts:
> Suppose you would do research on a very complex disease like asthma
> and hypertension. You know only some population frequencies and you
> expect a genetic component in some families - but you don't know any
> mode of inheritance. No strong candidate genes exist. How would you
> start a study? Collect families for the haplotype relative risk method?
> Look for affected sib pairs? Or do a linkage analysis in large pedigrees?
> Something else? All opinions are welcome.
> Regards, Matthias
>WJST at GSF.DE
I believe that, given our current techniques, one should (if possible) try
to find a set of pedigrees with extreme phenotypes related to the trait of
interested, preferably large pedigrees in which the phenotype is densely
represented. These are more likely to represent pedigrees for which a
mendelian model may be appropriate (and thus to benefit from the power
supplied with likelihood methods under an assumed mode of inheritance). It
is probably also wise to use more than one linkage analysis method in a
genomic search - both lod scores and, e.g., one of the methods related to
affected sib pair methods (including methods which included other affected
relative pairs, or which assume a quantitive phenotype). The specific
method(s) of choice will, of course, depend on the phenotype and type of
pedigrees in the data set.
In order to succeed in this endeavor, it is also imperative to have
sufficient resources and strong collaborators who are knowledgeable about
the different aspects of the project. The analysis techniques, the
molecular techniques, and even often the definition/description of the
phenotype are all in a great state of flux for complex diseases. Thus it
is necessary to have a strong clinician, molecular biologist, and
statistical geneticist on the project, at a minimum, all of whom have a
real commitment to the project.
Ellen Wijsman
Div of Medical Genetics, RG-25
and Dept of Biostatistics
University of Washington
Seattle, WA 98195
wijsman at u.washington.edu
