Abstract

For diffractometer data collected using a step-scan method, various procedures have been proposed by which the integrated peak intensity may be estimated from the measured reflection profile. However, these all ignore some of the information available in the data, thereby reducing the accuracy of the estimation. Moreover, some make assumptions about structure present in the sequence of counts and so produce a large positive bias in their estimation of weak refections. A profile-fitting approach based upon the Bayesian three-stage regression model is presented, which avoids these failings. The underlying theory is discussed, its implementation for off-line data reduction and its potential for on-line diffractometer control is described and its application to various protein data sets collected using both single- and multiple-counter diffractometers is reported.