Introduction

For almost 20 years we have been working on improving the quality of PDB files. Initially this work has mainly consisted of determining what is wrong in PDB files (and believe me, that is a lot). But for a couple of years already, we are working on corrections of PDB files. We started with the correction of NMR files simply by recalculating them from the original NOE data. And recently we have started working on the recalculation of X-ray structures from the original reflections.

Although we have written many articles about the quality of structures in the PDB, we still occasionally encounter people who think that "It isn't that bad, ain't it?", or who think that all errors are simply the result of the experimental nature of solving macromolecular structures.

This web-site is going to be a nicely illustrated demonstration of all types of errors that one can encounter.

Throughout these pages I will often illustrate things using (manually modified versions of) crambin. This does not say anything about the quality of this molecule, but only that I like its simplicity. The reference for 1crn is:

REMARK   1  AUTH   M.M.TEETER                                           1CRNC  3
REMARK   1  TITL   WATER STRUCTURE OF A HYDROPHOBIC PROTEIN AT ATOMIC   1CRNC  4
REMARK   1  TITL 2 RESOLUTION. PENTAGON RINGS OF WATER MOLECULES IN     1CRNC  5
REMARK   1  TITL 3 CRYSTALS OF CRAMBIN                                  1CRNC  6
REMARK   1  REF    PROC.NAT.ACAD.SCI.USA         V.  81  6014 1984      1CRNC  7

Throughout these pages the reference for each structure mentioned will be copied directly from the PDB file and shown in a box, just like this one. More generally, text in boxes comes (more or less) directly from the PDB files by cut-n-paste. Often, when larger parts are cut-n-paste from a PDB file to provide extra evidence to the error, so-called supplemental material has been used. This is shown as:

One day, perhaps, I will bring some order in the listing of error-types. For now, these pages are simply an illustration of everything that can go wrong while solving NMR or X-ray structures. Consequently, all sections are organized by PDB-id. This organization causes the occasional error to end up in the wrong section, but it allows you to visit any PDB file just once.

Nowhere do I try to be complete. Most errors that I list in these pages occur many, many times in the PDB. I just list those where the error is particularly clear, or sometimes I simply display only the PDB file in which I found the error for the first time.