| Gene imprinting is thought to a?ect seed development in plants. However,a statistical method that can detect and estimate the e?ects of imprinted quantita-tive trait loci (iQTL) throughout the genome has not extensively been developed. Wehave proposed two imprinting models in reciprocal backcross design for testing andestimating the e?ects of iQTL that determine variation in a quantitative trait. Theexperiments were designed using reciprocal backcross generations, so the inheritance ofparent-specific alleles could be traced. The computing model and algorithm were im-plemented with the maximum likelihood approach and EM algorithm. The imprintingmodels we proposed allow for a genome-wide scan for the existence of iQTL and testsfor their parent-of-origin e?ects. Monte Carlo simulation studies which include severalrepresentative schemes with di?erent imprinting degrees (small and large), heritabili-ties (0.1 and 0.4) and sample sizes (100 and 400 for each backcross) were performed toinvestigate the statistical properties of the imprinting models for the data simulatedunder di?erent imprinting degrees. The false positive rate (4%-7%) of iQTL discoveryby the imprinting model was examined by analyzing the simulated data that contain noiQTL. In general cases, the imprinting model covers the Mendelian model, suggestingthat one can safely use the imprinting model to analyze data, regardless of whetherthere is an iQTL. The imprinting models presented were applied to study the mode ofinheritance for iQTL that control endoreduplication traits in maize endosperm, lead-ing to the identification of a few iQTL that trigger maternal parent-of -origin e?ectson two measures of endosperm endoreduplication–mean ploidy and the percentage ofendoreduplicated nuclei. The reciprocal design and a series of analytical and testingstrategies we proposed provide a standard procedure for genomic mapping of iQTLinvolved in the genetic control of complex traits in experimental crosses.
|
PDF Full Text Request