In the mmCIF dictionary, _atom_sites.cartn_transf_matrix is described as the '3x3 matrix and used to transform fractional coordinates in the ATOM_SITE category to Cartesian coordinates in the same category and the following example is given: ; Example 1 - based on PDB entry 5HVP and/or laboratory records for the structure corresponding to PDB entry 5HVP ; ; _atom_sites.entry_id '5HVP' _atom_sites.cartn_transform_axes 'c along z, astar along x, b along y' _atom_sites.cartn_transf_matrix[1][1] 58.39 _atom_sites.cartn_transf_matrix[1][2] 0.00 _atom_sites.cartn_transf_matrix[1][3] 0.00 _atom_sites.cartn_transf_matrix[2][1] 0.00 _atom_sites.cartn_transf_matrix[2][2] 86.70 _atom_sites.cartn_transf_matrix[2][3] 0.00 _atom_sites.cartn_transf_matrix[3][1] 0.00 _atom_sites.cartn_transf_matrix[3][2] 0.00 _atom_sites.cartn_transf_matrix[3][3] 46.27 _atom_sites.cartn_transf_vector[1] 0.00 _atom_sites.cartn_transf_vector[2] 0.00 _atom_sites.cartn_transf_vector[3] 0.00 This would seem to imply that the coordinates are given in fractional. But in the atom_site section, the coordinates are given in an orthogonal coordinate frame. My first question is whether this example is correct. I then looked at the examples given at Rutgers. The coordinates also are orthogonal in each of the four examples. _atom_sites.frac_transf_matrix is the transformation that one would expect to find to make fractional from the orthogonal coordinates in the entry. _atom_sites.cartn_transf_matrix is an identity transformation in each of the four examples. My second question is why this transformation is an identity transformation in each of the four examples when it is not an identity transformation in 5HVP. Frances C. Bernstein