Same set of molecules at a collection of different heights

Challenges

• Stage tilted: all molecules in a given frame DO NOT have the same wx wy distortion.
• Stage started position uncertain — might not be mid-range.
• Stage z-drift during movie acquistion is not corrected.

Approach

• Use current estimate of z to compute stage tilt. This + stage position gives the best guess True-Z of every molecule in every frame.
• should use output of where pizo stage actually goes, (including the drift, which means the start and end frame are not the same). In the .off file this is the 2nd column, not the 4th column, which is what S-H S.’s code uses.

Should be able to just look at this plot, see that the current estimated angle is incorrect, attempt a different angle on the stage tilt, and get a much better fit to iterate on. Will try this tomorrow.

(here I have actually traced molecules through consecutive frames and color coded individual molecules. The individual traces are very sharp, good polynomial curves. They don’t all line up because they aren’t all running through the same range of z-values (some start further in or out of the plane than others). If we correctly compute their true positions this tilt means we get free data for a larger range of z-values to do the fitting on. But correcting that tilt means knowing the z-displacement in the first place, hence the iteration.

Maybe try discussing this general issue with XZ next week — no lab standards for data processing. “Manual chromatic correction,” vs bead fields before and after every imaging session. Lack tools, lack discussion.