Recombination plays a vital role in the proper disjunction of chromosomes in meiosis and is essential to the production of viable gametes. Many organisms use a strategy called interference to regulate the distribution of crossovers among all recombination events that include simple gene conversions (non-crossovers) as well as crossovers. Recent evidence suggests that the organisms that also use recombination to pair their chromosomes early in meiosis have an additional set of crossovers that is not subject to interference. The investigator develops the statistical theory to test this two-pathway hypothesis by extending previous techniques used to model the distribution of crossovers subject to interference. The investigator analyzes data from a variety of organisms to determine if the two-pathway model provides a better fit than previous models to recombination data from organisms that use recombination to pair their chromosomes during meiosis. Multivariate optimization techniques contribute to the development of tools for determining the design of experiments to test this hypothesis and to obtain sharp estimates of the model parameters. The investigator explores the advantages of using the two-pathway hypothesis in gene mapping methods by comparing its performance with simulated data to that of methods using no-interference and interference-only models of recombination.
A basic understanding of recombination is important for understanding certain causes of infertility, miscarriages, and birth defects in humans. For instance, trisomy is related to certain exchange configurations on the chromosomes during meiosis being susceptible to missegregation (for example, an egg receives an extra copy of chromosome 21 and a child resulting from the fertilization of this egg has 3 copies, two from its mother and one from its father, leading to Down syndrome). Trisomy is involved in up to 25% of miscarriages in addition to birth defects. Furthermore, mathematical models for the distribution of exchanges between maternal and paternal chromosomes in meiosis provide the fundamental basis for locating genes, including those that are linked to genetic diseases. Including the most appropriate models for these exchanges in gene mapping algorithms improves their efficiency without increasing the error rate they claim. If the proposed model proves to fit data the best, its inclusion in gene mapping algorithms will become increasingly important due to the use of Single Nucleotide Polymorphisms that provide a large number of markers for finding genes associated with complex disease traits. Additionally, the proposed research intimately connects the statistical methods to a current, empirically testable, biological hypothesis about recombination. The results will be disseminated to both the statistical and biological communities and the activities supported by this grant will include training graduate students in this interdisciplinary area.
