Thursday 06/19/14

10:00 am – 8:15 pm

Working on Python OpenMM simulations

• goal, attempt to execute some of Mirny’s example code.

Attempting to run demos

• e.g. scaffoldExample.py
• complains about notfinding joblib. installed this, required instlaling
• this complains about not finding mirnylib, which I downloaded along with mirnylab and put both folders independently inside Lib\site-packages.

Attempting to install Cython

• errors
• This is kinda stupid that it reports setup.py not found given that it is running the file called setup.py
• package directory ‘Cython’ does not exist

Attempting to install mirnylab\setup.py install

• same errors

The usual joys of python:

• Endless package dependencies on package dependencies. Many of which are non-standard (e.g. mirnylib, mirnylab, mirnylib numutils_new) and which neither validate their dependencies when you install them nor go and fetch these dependencies on the fly when they encounter errors.
• It seems to me this would not be too hard to fix — a missing import package is a pretty clear error, and it could direct to a download weblink.

Branching Bogdan’s polymer sim

• at least this code executes (on Monet as well as Tuck, after a few installs yesterday).
• modified attractive force (with step functions to enforce short range behavior) and other parameters taken from Minry’s simulations seems to largely solve the exploding polymer problem. Occassional small bursts are still observed. Hopefully these aren’t frequent enough to seriously disrupt the statistics.
• initial simulation (single run) looks promising for sub volumetric scaling.

Matlab Based simulations

• running fast scale test using multi-threading locally on Monet
• wrote script generator to dispatch jobs on Odyssey

Sequential Hybes

Observations

• by movie 26 (22?) they COT has switched to the keep-everything-on mode = very high background, especially from nonattached probes.
• Running bleaching
  • first round of bleach movies in 2X SSC (no STORM buffer) still observe pretty decent switching and little spots in previously imaged (‘to-exhaustion’, obviously not true) spots. (10,000 frame movies)
  • even with minimal or no 405, substantial bit of switching.
  • Should probably image all these sequential samples without COT. previously reached this conclusion on project2
  • running second round of bleaching, 10,000 frames. Moved original bleach movies to backup to avoid overwriting.
  • bleaching with 1270 mW 647 (STORM with 270 mW)
• hot wash at 37C in 50% formamide in 2X SCCT 1 hour
• wash further 20 min in 2X SCC (remove formamide)
• STORM now with no COT.
• pretty clear, 1 hour wash has no free dyes floating around now, unlike first wash. (maybe use less foramide?)
• forgot to drop power for first movie

Previous F03F04 results

• with 750 secondary, with drift and chromatic correction
• good alignment
• low localization count