Computational alchemy apprentice
For the last few months, I’ve had the great pleasure to work in a chemistry lab at McGill university alongside highly skilled and motivated scientists. I am not a chemist, so my job was simple: re engineer a molecular simulation program and make it run faster than the original using GPUs instead of CPUs as processing units. One of the great haha moments came when my collegue Ata and I managed to move our first molecule inside a protein using the conjugate gradient minimization method. At the time, I was only inspecting decreasing energy numbers dumped on a black terminal, but I used Pymol (an Open Source molecular visualization program) to create the above movie.
What is shown here is a graphical model of the Herpes Simplex Virus Type 1 Thymidine Kinase Interacting With (North)-Methanocarba- Thymidine. The kinase is a large protein present in the body (shown as gray dots representing atom positions, and gray lines representing bonds), and the colorful molecule is the “cure”. The movements are rather small and jerky: they are the result of successively applying movements on each colored atom in the direction of the forces computed by our software.
But my adventures in Bio computing are coming to an end: I am taking my GPU skills to new adventures as of next month.
stay tuned…
http://en.wikipedia.org/wiki/Conjugate_gradient_method
http://en.wikipedia.org/wiki/Force_field_(chemistry)
http://en.wikipedia.org/wiki/Molecular_modeling_on_GPU

