DDQ version 3.0 (9/Sep/2026)

To use DDQ, first, delete all the water molecules from your structure and run a quick crystallographic refinement (using for example Refmac) to generate new coordinates and a new .mtz file. Next, use CCP4's FFT program to calculate a Fo-Fc map (nF1-mF2) and have it cover all atoms in the PDB file. Check the box for "Do peak search of map". Select F, PHIC_ALL, FC_ALL, SIGF, FOM, and unassigned for the F1, PHI, F2, Sigma, Weight, and Sigma, resp. Enter the name of your protein coordinate .pdb file. Also enter the name of the peaks coordinate as you will need that file to run DDQ. Enter n=1 and m=1 for the Fo-Fc difference map, use a 5 Angstrom border. Search for peaks greater than 3.0 sigma and also check the box "Search for negative peaks". 
To compile DDQ fortran code on Linux:
gfortran -O2 -std=legacy -ffixed-line-length-none ddq9sep2026.f -o ddq9sep2026
To run the program, type ./ddq9sep2026
(and make sure you have these four files in the same director: symmetry, angles, atom_rad, and param.)
Type in CCP4 for the first question, then the name of the peaks file, followed by the name of the protein coordinate .pdb file.

The DDQ program is provided free and on an "as is" basis. If you have questions, please contact me at focco.vandenakker@case.edu .

A few notes about this program: 
My old crystal structure validation program DDQ has now been ported to run on Linux, with the help of some code changes using ChatGPT. This crystal structure program probes for an almost impossible to fabricate fingerprint of a correctly refined crystal structure. After the water molecules are deliberately removed from the structure factor calculation, only well refined crystal structures will have: a) accurate crystallographic phases that aid in generating strong positive electron density peaks representing these missing water molecules; b) these water density peaks are situated at chemically sensible distances and angles from hydrogen bonding atoms (and non-hydrogen bonding atoms); c) an absence of negative density shift peaks on top of atoms if they were incorrectly modeled; and d) absence of positive density shift peaks near atoms where one should have modeled atoms. These features are all incorporated into a single number that is compared to other structures in the PDB for the structure's particular resolution bin giving you a resolution-dependent ranking of your structure (best, top 25%, above average, below average, bottom 25%, worst). To my knowledge, DDQ was the first crystal structure validation program that used diffraction data to aid in giving resolution dependent ranking for crystal structures. The DDQ program is also very sensitive in finding local errors in crystal structures or assessing whether local regions, including ligands, are modeled well. The original publication describing DDQ can be found here: https://pubmed.ncbi.nlm.nih.gov/10089411/


-Focco van den Akker
