Multiple changepoint fitting via quasilikelihood, with application to DNA sequence segmentation

We consider situations where a step function with a variable number of step s provides an adequate model for a regression relationship, while the varia nce of the observations depends on their mean. This model provides for disc ontinuous jumps at changepoints and for constant means and error variances in between changepoints. The basic statistical problem consists of identifi cation of the number of changepoints, their locations and the levels the fu nction assumes in between. We embed this problem into a quasilikelihood for mulation and utilise the minimum deviance criterion to fit the model; for t he choice of the number of changepoints, we discuss a modified Schwarz crit erion. A dynamic programming algorithm makes the segmentation feasible for sequences of moderate length. The performance of the segmentation method is demonstrated in an application to the segmentation of the Bacteriophage la mbda sequence.