A new statistical model for describing errors in isomorphous replacement data: the case of one derivative

A probability method for phasing macromolecular isomorphous replacement data from the viewpoint of errors arising in the isomorphous replacement process is considered. The assumption of imperfect isomorphism between the atomic positions in the native crystal and the atomic positions in the native component of the derivative crystal leads to estimates of phase dependent on the value of sin θ/λ. The mathematical techniques used are similar to those employed in deriving probability distributions of structure seminvariants. Some of the formulas, which apply only to the case of one derivative, are compared with earlier results.