Tips and Tricks in Crystallography

Space group validation with the Zanuda program in CCP4.

Dr. Mohana Rao (NCI): "Headache" Medicines. The X-ray data are collected, indexed, and processed. The R-merge is good. But still one feels a bit queasy about it. Two cell dimensions are close and the angles seem to be funny. Maybe the crystal system chosen by Denzo is not the right one. Maybe there is a centering somewhere. I am presenting a simple way to study these. During the last week or two, I had the chance to discuss at length issues like these not once but twice. The first thing I did was to create a coordinate file (WH type) that has lattice points from -3 to +3 in all directions. Then these were converted to a PDB file using the (doubtful) cell dimensions. The lattice points were then displayed using a program like Insight. A typical cell was constructed by joining the various lattice points. Viewing this in three dimensions and rotating it in all possible ways, it was possible to obtain alternate cells- from one triclinic to another triclinic cell in one case and from monoclinic to C-centered orthorhombic in another case. In this way, not only the correct choice of the unit cell may be made, but also how the whole thing fits in three dimensions may be understood. However, these have to be later confirmed by rescaling the data with proper transformation matrices.