Tuesday 02/19/13

10:00A – 1:40P, 3:30P-12:10A

Computer

  • restart required for hotswap and to switch Alistair8 back to eSATA
  • delete HB data from Alistair3b (now on Alistair3).
  • disconnect Alistair3
  • future disk shuffling of earlier data might be better done on Tuck?
  • remove Buffalo TurboCopy (this is def. slower than windows copy, at least compared to network to USB vs tubo USB to USB).

Fly work

  • collect AM Esc virgins
  • set up new collection cages of 29C MAT-alpha tub fly lines: Rad21, Scm, Psc1, and Pc3.
  • moved Mat tub VP16 on 3 crosses to 18C virgin collection

Embryo work

  • flip collection cages, 12P
  • fix Esc embryos 6-18 hrs.
  • flip collection cage ~6:30P.
  • fix Esc embryos, 4-10hrs
  • make new embryo collection cages

STORM analysis

  • Splitting large multifaceted clusters into multiple 3D gaussians is sorta what 3D DaoSTORM is doing. We should be able to call this command directly. Except we dont’ wan’t to assume that each blob is made up of potentially different number of otherwise equal gaussians.
  • I guess we really just want the first step of multi-emitter — fit a gaussian to every local maxima.
  • Now to write¬†a code to find local maxima in 3D space… scan through threshold from 0 to 1, maximize number of objects in image?
  • “waterfilling”, call each peak as it vanishes.
  • should REALLY check google / stackoverflow first: http://stackoverflow.com/questions/1856197/how-can-i-find-local-maxima-in-an-image-in-matlab
  • Finally have a more robust subcluster segmentation algorithm (I hope): findclusters3D.m (function). Still need to run through BXC script.
watershed3D_clustering

watershed3D_clustering

more_elegant

more_elegant

findall_local_max

findall_local_max

ubx_cluster

ubx_cluster

guassian3D_clusterstats

guassian3D_clusterstats

This entry was posted in Summaries and tagged , , , . Bookmark the permalink.