autotrack-roadmap.rtf

{\rtf1\ansi\ansicpg1252\cocoartf1265\cocoasubrtf210
\cocoascreenfonts1{\fonttbl\f0\fswiss\fcharset0 Helvetica;}
{\colortbl;\red255\green255\blue255;}
\paperw11900\paperh16840\margl1440\margr1440\vieww22640\viewh18600\viewkind0
\pard\tx566\tx1133\tx1700\tx2267\tx2834\tx3401\tx3968\tx4535\tx5102\tx5669\tx6236\tx6803\pardirnatural

\f0\fs24 \cf0 Roadmap for the autotrack software:\
\
- Load project : \
\
simple command line interface to load project and set analysis parameters \
detect if data are already segmented using autotrack DONE\
\
stores param\'e8tres in global autotrack variable DONE\
\
%autotrack.position\
%autotrack.path\
%autotrack.parameters\
- test processing data : determine appropriate threshold level, etc\'85 display results obtained for specific frames DONE\
\
- Batch engine to process data\
loop on positions, segment, and map objects, save segmentation variables DONE\
\
MAPPING : add fluo intensity to fix some M/D assignment errors. \
\
-log file generator DONE\
\
- check lost cells and mapping errors : Camille function to check DONE\
\
- pedigree analysis to determine parentage and list errors : PEDIGREE DONE, BUT NOT USED TO CHECK ERRORS\
\
- use nucleus mapping to perform cell mapping backwards TO DO\
\
-Fit gaussian to determine true flu level DONE\
\
-Extract cell cycle phase\
TO DO : \
*separate detection of division from cell cycle phase  ( 1 parameter to add) : example possibility to track div times in HTB2-mCherry DONE\
\
*div time should be defined as end of anaphase (and not middle of anaphase) : definition based on minimum value obtained for nucleus fluo DONE\
\
* dual fit to extract both ramp up  (S phase) and ramp down (anaphase) DONE\
\
- GUI to plot results and load raw data from scatter plots DONE\
\
- data analysis program\
\
the program should be modular so that data could be checked and errors detected  at each step.  DONE\
\
\
To do : filter to discard wrongly fitted data  DONE\
\
- time filter DONE\
- G1 and post anaphase fluo level should be in the same range DONE\
\
\
- make movie corresponding to ne cell cycle\
\
\'97\'97\'97\'97\'97\'97\'97\'97\'97\'97\
\
Strategy to add phase contrast image // cell size : \
\
segment and mp nuclei  - potential tracking error due to M/D swapping upon nuclear division\
quantify nuclei\
\
new ideas : \
\
- Hungarian method (cf JL Tynevez) : use (M+N) x (M+N) matrix to add a cost associated with lost or new tracks, may be better than using an infinite cost if cell moves too much\
- cavity : prevents cells from going down ; can olnyl go up; 2) a cell cannot disappear , unless it dies ! \
- add a model to anticipate cell motion\
- establish links between groups of cells : conservation of polygon patterns / use accumulator\
- closing gaps in trajectory ; this is done using optimization of a cost matrix in JY Tinevez algorithm\
\
- MHT tracking , JPADF fitler, particle filter for multi target tracking\
\
- Youlian : us cell mapping that is quite robust to help determining the pedigree ; how to combine information in the best way is tricky \'85. DONE !\
\
\
March 2014\
\
- added volume area + growth rate DONE March 2014\
- use cell mapping as a starting point for trajectories DONE March 2014\
\
\
TO DO : \
- improve stat and display tools  : trajectories DONE\
- analyze and compare distributions for different datasets DONE\
- generate movies for cells of interest on the fly\
- generate movies for typical cells corresponding to a specific mutant \
- increase acquisition throughput using nikon microscope DONE\
DATA : \
\
mutant data\
effect of maturation time  TO DO analysis\
comparison of nucleus scoring methods TO DO analysis\
\
\
Sandrine analysis \
ROI on cavities  DONE\
extract M/D divisions from multiparameter curve : cavity.m DONE\
\
\
\
\
\
\
\
\
\
}