User:Steven J. Koch/Notebook/Kochlab/2010/05/03/Tracking 80% D2O, 81%O-18, and 0% (alpha)...peaks!

Today, I used the manual version of Larry's tracking software to help Andy with his talk. I tracked the following (just a tiny fraction of his data!):


 * 10th (and 9th?) file for 80% D2O from April 28th.
 * 10th file for 81% oxygen-18 water from April 29th.
 * 10th file for alpha casein from March (30?) (this is the 0% control).
 * _most_ (i forgot some) of my analyzed microtubules can be found in this public directory: http://kochlab.org/files/data/Microtubule%20Tracking%20Data/2010%2005%2003
 * I need to tweak the software so the public copies are made automatically. There's some tricks, though, since we sometimes make mistakes adding to the array.
 * _all_ of the tracking results (including trials, errors, etc.) are still on the private server.

Main solid result: Oxygen-18 water has a significant effect on gliding speed! Not as much as D2O, but very significant, I would say. This could mean that the effect of both of them is not due to isotope effects directly relating to deuterium bonds. Oxygen-18 water has protons, and thus many of the crazy effects of D2O are absent. Viscosity, I think, is unlikely the answer, since oxygen-18 water isn't much more viscous than regular water (I think 5%), whereas D2O is 22% more viscous. It's possible that we really are probing the "hydration" of the motor domain / tubulin interaction areas, but we're a long way from proving that.

Other result: From perspective of looking at data, I am convinced that the microtubules have distinct speeds that they travel at, seemingly unrelated to length. I have also seen data that look convincingly like there was a switch in speed for a given MT. When I selected all the most-likely speeds from individual speed PDFs for individual MTs, all three conditions looked strongly bi- or tri-modal. This seemed obvious and kernel density estimation analysis of the peaks of the PDFS looked great. However, when I tried doing KDE on all of the speed data together, it was more difficult to see the peaks. I leave it to my evernotes below for my current thinking. Currently, I think the peaks when using a bandwidth of 4 nm/s are real.

Related friendfeed thread
Neil Saunders commented on my automatic friendfeed entry. 

Evernotes
  