User:Karsten Theis/Electron density

To explore the electron density of the structure 233D (a short piece of DNA containing some sulfur-modified sugars), choose one of the two scenes below (either the entire molecule, or just a base pair. The latter will load faster and is less confusing, so maybe start with the base pair).

The following is the command to fetch the electron density from the pdbe.org database and display it around the selected atoms:

isosurface s_one sigma 1.0 color blue within 1.5 {selected} "http://www.ebi.ac.uk/pdbe/coordinates/files/233d.ccp4" mesh nofil

Electron density

Click on the following buttons to display electron density at different contour levels. One sigma means to show the area (within the chicken wire) that has values higher than one standard deviation over the mean. The higher the sigma, the less density you see (but you get to see the highest peaks in the density).

The peaks are mostly around the phosphorous atoms. P has about twice as many electrons as C, N and O, so it is expected to have roughly twice the electron density (and, even more roughly, twice the maximum electron density at the center of the atom).

At 3 sigma, you begin to see other parts of the structure when it is highly ordered. You also see the sulfur atoms, which are also third row elements.

This is a good contour level for modelling. You can see evidence for most of the atoms, but the level of noise is not that bad.

If you don't see your model at 1 sigma, you probably should not put in those atoms. At this level, you see a lot of solvent molecules and a lot of noise.

A technical note

This is not the most efficient way of displaying electron density, and it works because the structures displayed here are rather small. Every time you click on a button to display electron density, the entire electron density file is fetched from the database, and your browser has to read it all to find those points in space that are close to the selected atoms and have a density higher than the cutoff.

Update: there is now an efficient way to just fetch the part of the map you need:

isosurface within 3 {visible} "=233D"

will retrieve that part of the map within 3 Angstroem away from any atom displayed.

Try it:

If you already know what density you want to show, it is much faster (and saves space) to save the "chicken-wire" object as a .jvxl file and upload it to Proteopedia. These files are much smaller (especially if you are showing the density around just a couple of atoms, which is the typical case for publication figures). An example of this technique is explained here.

Ferrodoxin: Ultra-high resolution

is a small protein that binds to a certain metal.

For this structure, we can look at the density at very high contour level to find the metal.

This will take long to load... the density file has a size of 11.2 Mbyte and your browser might ask you whether you want to continue. Once its loaded, see it you can spot the metals in the structure. To see the interpretation, click the button below.

If your patience did not run out, you can look at the density at the typical 1.5 sigma to see all atoms.

Pre-calculated density

This is again, surrounded by 20 sigma density. It loads faster because the density was rendered already and saved as a small file (10 kB instead of 11.2 Mb - factor 1000 less).

Even the 1.5 sigma density around loads faster (222kB file). This is the kind of display you would use to build or rebuild an atomic model into electron density. To be able to see the trees in the forest, you use a technique called slabbing, illustrated below.