mjo22 / cryojax

Cryo electron microscopy image simulation and analysis built on JAX.
https://mjo22.github.io/cryojax/
GNU Lesser General Public License v2.1
26 stars 9 forks source link

More realistic water model #236

Open geoffwoollard opened 2 weeks ago

geoffwoollard commented 2 weeks ago

Make a cheap model of water placement (Monte Carlo?). This could build off of a routine in openmm that places water and removes them if they are too close (a type of rejection sampling).

https://github.com/openmm/openmm/blob/51a112a336db6cc0aea1c32b90d9d0a1f262cbfd/wrappers/python/openmm/app/modeller.py#L384

geoffwoollard commented 2 weeks ago

@mjo22 do you think this is good functionality to have?

I think it would go well with relaxing the projection assumption (high levels of scattering theory) and non-gaussian noise (e.g. Poisson).

geoffwoollard commented 2 weeks ago

Possible model to implement Shang, Z., & Sigworth, F. J. (2012). Hydration-layer models for cryo-EM image simulation. Journal of Structural Biology, 180(1), 10–16. http://doi.org/10.1016/j.jsb.2012.04.021

It’s an implicit solvation model. A way to evaluate the 3D potential from atomic coordinates, in a way that takes into account the solvation effects.

The code is available here: https://www.mathworks.com/matlabcentral/fileexchange/33636-cryo-em-image-simulator

mjo22 commented 4 days ago

Changing the name of this issue since it is not yet decided how to approach an improved model of the solvent