Standard approaches to linkage mapping of complex traits such as diabe
tes encompass two basic techniques or strategies: 1) tracing of cosegr
egation and recombination phenomena between marker alleles and putativ
e trait-influencing alleles among members of large pedigrees (i.e., pe
digree analysis) and 2) assessing the number of marker alleles shared
at a particular locus among pairs of relatives manifesting the same tr
ait (sib pair or allele-sharing analysis). These techniques rely on di
fferent assumptions and statistical models and thus are appropriate fo
r different situations. In addition, Each technique possesses certain
advantages and disadvantages: pedigree analysis is often assumption-la
den but powerful, whereas allele-sharing methods require few assumptio
ns but are not very powerful. In this review, we discuss the relative
advantages and disadvantages of pedigree and allele sharing analyses,
as well as the possible extensions of these techniques that might make
them more flexible. Ultimately, however, we argue that hybrid statist
ical models and methods that draw on aspects of each approach are like
ly to be the analytic tool of the future for mapping genes that underl
ie complex traits.