Newsletter 161
December 3, 2007

The NIH X-Ray Diffraction Interest Group

Newsletter web site:

Item 3: Topic Discussion - Twinning - (Click for PDF reader to view articles)

Please share your experience in twinning by presenting a case study. To learn or to review the basics of twinning, see CCP4 General and Paul Adams' presentation.

Dr. Xinhua Ji (NCI): Perfect Merohedral Twinning - No News Is Good News  
      Perfect Merohedral twinning involves a crystal with a twin fraction equal to ½. One of our recent crystals was perfectly twinned in space group P43. Nonetheless, MAD phasing and structural refinement were carried out in space group P43212 until the refinement stuck at Rfree = 0.38. The electron density was outstanding for the resolution and no additional features were revealed by phase combination. The refinement was then continued and finished in space group P43 with perfect twiining.
     We knew the possibility of perfect twinning, but we did not bother until we had to. This "no news is good news" approach offered three advantages. First, there was one molecule in the asymmetric unit of P43212, but two for P43. Therefore, the phase problem without twinning was half of its actual size. Second, there were eight equivalent positions in the unit cell of P43212, but four for P43. Thus, the redundancy of observations doubled without twinning . Finally, model building and adjustment were easier working with one molecule. Since no reliable means is available to detwin diffraction data from perfectly twinned crystals, it is not beneficial to treat twinning before it is necessary to do so.

Dr. Mark Mayer (NICHD): A Narrow Escape from Merohedral Twinning 

    Merohedral twinning is a special form of disorder that most crystallographers will be forced to deal with at some point in their career. It is different from twinning which arises when crystals fuse during growth, which is easily recognized either in the light microscope, or from diffraction images which reveal the presence of more than one lattice. In the case of merohedral twinning, the crystal contains microdomains in which the same lattice is present but in different orientations related by a twinning operator. As a result, the observed intensities are not accurate, in the sense that they arise from the sum of the unrelated intensities of the twin components. Because the intensities do not correspond to those generated by a single lattice, refinement stalls at unreasonably high R values, and in some cases the structure cannot be solved at all. (Full Article)

Dr. Lothar Esser (NCI): Refinement in Case of Twinning 

    Despite working in crystallography for a number of years, it was only recently that I was asked to refine a structure against twinned data. The unfinished structure was handed to me after the post-doc who had been working on this project had left the lab. The crystal diffracted x-rays to 2.2 Ǻ with intensities obeying the symmetry of space group R3. The structure was solved by molecular replacement before it was recognized that the crystal was twinned using a model of > 99% sequence identity. As the refinement did not progress as it should, it became clear that the data were twinned. Merohedral twinning in R3 is quite common and the extra 2 fold axis, that generates the twin domain, relates reflection indices of h, k, l to k, h, -l. Having no prior experience with the refinement of structures against twinned data, I tried all programs that were available to me.(Full Article)

This site is maintained by Dr. Xinhua Ji ( on the NCI-CCR-MCL server (