Statistical Analysis of Ordered Tetrads and Half-Tetrads
Ordered tetrads and half-tetrads are two types of genetic
data that arise both from experimental organisms (e.g.
yeast, mold, fly, mouse) and humans (e.g. trisomy 21 which is
responsible for over 95% of Down syndrome cases and
uniparental disomy that may cause Prader-Willi syndrome
and Angelman syndrome). Although, from as early as 1920's,
both types of data have been studied by geneticists to
understand crossing-overs between homologous chromosomes
during meiosis, there has been no systematic study
on the stochastic processes involved in the generation of
these data. In this talk, I will first describe the biological
processes underlying ordered tetrads and half-tetrads, and
then discuss the assumptions usually made in analyzing these data.
We will show that these assumptions impose certain constraints
on the probabilities of different types of ordered tetrads
and half-tetrads. These constraints can be used to
test the common assumptions on the random processes
leading to recombinations and to order multiple genetic markers.
In addition, to facilitate the genetic mapping using ordered
tetrads and half-tetrads, explicit multilocus ordered tetrad
and half-tetrad probabilities will be derived under a class of
chiasma interference models, the chi-square models.
With real data sets, I will illustrate how to apply our
results to order genetic markers, carry out centromere mapping,
and study the distribution of crossover points during meiosis.
This is joint work with Terry Speed at UC Berkeley and Jinming Li
at Yale School of Medicine.
Seminar to be held in Room 107, 24 Hillhouse Avenue at 4:15 pm