Detection of Patterson ghost peaks and shape distortions due to sampling in Patterson functions from incomplete data sets

A procedure for evaluating peaks in the difference Patterson maps calculated from a complete native data set and partial (incomplete) potential isomorphous heavy-atom-derivative data sets is described. A sampling function is defined such that it is unity where data are available and zero elsewhere. It is the Patterson of this sampling function whose characteristics are useful. This function will not reconstruct unavailable information, but can reveal the presence of distorted and/or `false' peaks introduced by the sampling technique used in collecting partial data. This procedure is useful in screening potential heavy-atom derivatives using a minimum number of crystals.