I recently came across this excellent commentary by Brian Matthews on the 5 (five…) papers Chang et al. retracted back in 2006. For those not familiar with X-ray crystallography and the Changs papers withdraw from leading scientific publishers, I give you a bit of explanation. X-ray crystallography is the gold standard in structure determination and it uses a crystal of pure molecular specie(s) shot through an X-ray beam. If the crystal is good, the electron clouds of the atoms diffract the x-ray beam. By recording various diffractions images, one can compute the electron density inside the crystal and trace (build) the molecular specie(s) in it. Sound simple enough? actually no. I’m not talking here about the maths and physics involved and the phasing problem. The essence of  X-ray crystallography is to solve the phase problem leading to having “good” and “reliable” maps of electron density.

What happened to Chang et al., is that they were working with some wrong electron density maps because of a gross error made early-on during the project pipeline. The culprit was an in-house data reduction program that switched critical column of data. Because of this error, they build/trace various proteins with the “wrong” hand.

Brian Matthews commentary is a solid X-ray crystallography 101 lesson. A lot have been said and written about Chang et al. mistake and their consequences. But, Brian Matthews point out that nowadays we are seeing an ever-increasing use of “black-box” procedures for structure determinations. The rapid development of easy to use X-ray crystallography softwares along with massive computing power render the structure determination fairly easy for one with limited X-ray crystallography knowledge. Solving a structure can be fairly straightforward but it can easily become a tricky task. In any cases but especially in tricky cases, one needs to be extremely cautious about the validity of the maps. Brian Matthews gave us a great lesson about the various checking we all should do when dealing with problems encountered by Chang et al.

X-ray crystallography is like anything else, it is an art. It requires experience, failures, learning from failures and constant knowledge update. Like everything else in Science, it is a grave mistake to assume that we master all the whereabouts of a technology/methodology. I guess, Chang et al. learnt it the hard way. However it raises another question: Shall we seek advice from a peer to help solving a problem in case of dealing with a very hot project? All the 5 retracted papers were all hot projects…

The take home message from this is to be über-cautious and don’t take nice looking maps for granted…