I am impressed by the illustrations for the Protein Data Bank ‘Molecule of the month’, e.g. the gorgeous image of DNA Helicase below. Does anyone know how they were made or how one might create something similar?

DNA Helicase
(source: rcsb.org)

  • $\begingroup$ While I don't know what is used, David Goodsell should know it, as he is responsible for this, as far as I can see. Write him a friendly email, my best guess is that he will tell you. And please tell us (you can answer your own questions), I am curious, too. $\endgroup$
    – Chris
    Jul 2, 2014 at 16:34
  • $\begingroup$ Me also don't find it cool or impressive, However, Chimera has a visualization preset that makes this image style, just a click of buttons. In the presets menue click on publication 1 and BOOOM here you go the image will look similar like this $\endgroup$
    – Amr
    Dec 13, 2023 at 7:46

2 Answers 2


Those (really cool) pictures are created by David Goodsell using custom-written software.

From an interview to the artist:

PDB: How do you create the illustrations?

Goodsell: Most of the pictures are created with a computer program that I developed back when I was doing postdoctoral work with Dr. Art Olson here at The Scripps Research Institute. I've been using this style of illustration--with flat colors and black outlines--for about 10 years now. I like the way that this style simplifies the molecule, giving a feeling for the overall shape and form of the molecule, but at the same time you can still see all the individual atoms. On the last page of each Molecule of the Month--"Exploring the Structure"--I always use RasMol, to give visitors an idea of the kinds of pictures that they can create themselves with off-the-shelf software.

There are good tools around if you want to replicate that look. Although it may take a bit of tweaking (and possibly programming) I would surely give PyMOL and Bioblender a try.

  • 2
    $\begingroup$ I found QuteMol got me closest to Goodsell's look :) $\endgroup$
    – nornagon
    Jul 5, 2014 at 4:36
  • $\begingroup$ @nornagon: nice, I did not know about QueteMol! $\endgroup$
    – nico
    Jul 5, 2014 at 9:33
  • $\begingroup$ "I always use RasMol, to give visitors an idea of the kinds of pictures that they can create". Oh no they can't! Rasmol does not run on Mac OS X. But you can use Jmol instead, which has a command for saving high res png files that you can actually drag back into the application and modify further. $\endgroup$
    – David
    Feb 22, 2019 at 23:50
  • $\begingroup$ @David there are Mac OS X binaries for RasMol, although I have never tried them. rasmol.org/software/RasMol_2.7.3.MAC/index.html $\endgroup$
    – nico
    Mar 2, 2019 at 12:17
  • $\begingroup$ @nico — Not quite. They are Unix binaries that will run under an X-windowing environment. It requires you to set up Unix directories and move stuff there. I tried it years ago and it was a real pain compared with the Mac OS9 version. Now a user would have to install XQuartz before they started. For practical purposes it's dead. One feature of JSmol is that it supports all RasMol console commands, so if you are used to RasMol... $\endgroup$
    – David
    Mar 2, 2019 at 14:33

I don’t actually consider these pictures either “gorgeous” or “cool” — they are not to my taste — and I’m not really sure that the question is about biology, but when it resurfaced after almost 5 years I thought I’d give an answer which explained how one might create something similar, rather than how they were actually made.

The original 3D-graphics program used was RasMol, but as that only runs on Windows, hasn’t been updated for years, and has no viable web version, I would suggest using JMol/JSMol instead. This is free, and you can either download the application and work with it as described in my answer to a previous question. Alternatively you can just find a web page with a JSMol window (such as a page used in my own teaching material) and use that with the Console available from the JSMol logo. So what I do is:

  1. Get the PDB file I want to work with from the Protein Data Bank.
  2. Drag the file into the JMol window, which causes the protein to load in wireframe mode, as in (A), below.
  3. Give a series of commands such as the following to get a spacefilling image (B) in colours other than CPK (you can specify the colours of the chains if you wish) and without the shiny light patches, which are fine for standard images, but will cause problems. One useful feature of JMol is that it allows you to generate high resolution images:
    restrict protein
    wireframe off
    select protein
    spacefill on
    colour chains
    set specular off
    write pngj 2000 2000 "my.png"

You can subsequently drag these back into a JMol window and make further adjustments, as long they haven’t been edited. (So do graphic work on a copy.)

  1. Next you need access to decent bitmap graphics application. If you are in a University someone will have a copy of Photoshop (old versions are fine), but otherwise you will have to make do with what you can find — GIMP is a free cross-platform altenative. I am almost certain that the original artist would have developed the style he uses with Photoshop. He says he wrote a program himself to automate it, but he could just have well used a Photoshop script (Action) in my opinion. Anyway, for occasional use, you can just do it interactively.

  2. The first thing is to get the outlines on the spheres and make the colours flat. The former is done with a filter that finds the edges, and the latter by posterizing (reducing the number of colours in the image). I used a filter called ‘Poster Edges’ to produce frame C. There are sliders that allow you to change the intensity of the edges and degree of posterizing.

  3. Finally to reproduce the rather washed out feel of the colours I used the ‘Exposure’ tool (Image > Adjustments > Exposure), but I could have used curves (Image > Adjustments > Curves) or, no doubt, other options.

Stages in "Molecule of the Month preparation"

And there you have the final image shown in D. It doesn't have quite the same posterization as the DNA Helicase illustrated in the question, but if you really wanted that you could probably achieve it after a little experimentation.

  • $\begingroup$ Four years on, and when this question came up again I discovered some typos and missing links. I've remedied them now. Please flag me if you find anything else wrong. $\endgroup$
    – David
    Dec 14, 2023 at 14:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .