|
We describe a method for the interval mapping of quantitative trait loci (QTLs) for a time-to-death trait when a considerable proportion of individuals fail to die. This work was motivated by a mouse cross for susceptibility to Listeria infection, in which 30% of the mice recovered from infection, and the other 70% died within 240 hours, with an average (standard deviation, SD) time-to-death of 106 (33) hours. We consider a single-QTL model in which a mouse with genotype g at the QTL has probability pg of recovering from the infection, and, if the mouse dies, its log time-to-death follows a normal distribution with mean µg and SD s. We describe the calculation of three lod scores: LOD(p), to test the hypothesis that the pg are equal; LOD(µ), to test the hypothesis that the µg are equal; and LOD(p,µ), to test the combined hypothesis that both the pg and the µg are constant in g. |