Data pathologies caused by effects such as diffraction anisotropy and translational non-crystallographic symmetry (tNCS) can dramatically complicate the solution of crystal structures of macromolecules. Such problems were encountered in determining the structure of a mutant form of Rab27a, a member of the Rab GTPases. Mutant Rab27a constructs that crystallise in the free form were designed for use in the discovery of drugs to reduce primary tumour invasiveness and metastasis. One construct, hRab27aMut, crystallised within 24 hours and diffracted to 2.82 Å with a unit cell possessing room for a large number of protein copies. Initial efforts to solve the structure using molecular replacement by Phaser were not successful. Analysis of the dataset revealed that the crystals suffered from both extreme anisotropy and strong tNCS. As a result, large numbers of reflections had estimated standard deviations much larger than their measured intensities and their expected intensities, revealing problems with the use of such data at the time in Phaser. By eliminating extremely weak reflections with the largest combined effects of anisotropy and tNCS, these problems could be avoided, allowing a molecular replacement solution to be found. Lessons learned in solving this structure have guided improvements in numerical analysis used in Phaser, particularly in identifying diffraction measurements conveying very little information content. The calculation of information content could also be applied as an alternative to ellipsoidal truncation. The post mortem analysis also revealed an oversight in accounting for measurement errors in the fast rotation function. While the crystal of mutant Rab27a is not amenable to drug screening, the structure can guide new modifications to obtain more suitable crystal forms.