Marker difference regression, also known as joint mapping or simple marker
regression, regresses differences of means for marker classes ('marker diff
erences') on their expected value, which is a function of the QTL effect an
d position. The genome is scanned by a likelihood ratio statistic, which co
nditionally on the QTL position follows a chi-squared distribution. The unc
onditional distribution is a chi-squared process. This paper proposes a qui
ck method for controlling the genome wise Type-I error rate. The method is
shown by simulation to work satisfactorily for a number of settings.