Question: How does maths and computer coding help to correct the damage made by the X-rays?

Jonny Brooks-Bartlett answered on 23 Jun 2015:

When X-rays are fired at the crystals made of the proteins we get a spotty image which we call a diffraction pattern (bottom image on my profile page: /lutetiumj15-zone/profile/jonnybrooksbartlett/)

What we see in the experiment is that the spots start to fade. This isn’t good because we need to collect hundreds of these images to get a complete picture.

The intensity (brightness) of a spot on an image is just a number. The higher the number, the more intense the spot. If we can get the correct undamaged intensity number then we can solve the structure.

My project actually tries to estimate what number each spot should be. This is quite hard to do so I try to use statistics to get an idea of the correct value (the mean) and then I also try to provide an estimate of how far off I am (we call this the variance).
To go into slightly more detail, I try to work out how much the atoms in the crystal move due to X-rays hitting them, and estimate how spots change in relation to how far the atoms have moved. If you’ve come across something called “the normal distribution” I use that quite a lot.

So that’s the maths. I use computer code because there are too many maths equations for me to write by hand. I translate the maths into code and let the computer do the calculations for me.

The computer can also make pretty graphs for me like this one I made: https://github.com/JonnyCBB/StructureFactorCorrelations/blob/master/test/IntensityDiffComp_0-2-0_and_2-2-2.pdf