Wednesday, March 29, 2006

FRAP data importing

These two days working on..

- Igor analysis modules for the track coordinates resulted from new macro in ImageJ.

- made an interface for importing excel files to the FRAP analysis procedure (IgorPro), with seperated files each for background decay and FRA curve. (for Fatima)

Thursday, March 23, 2006

Aurelien's Image J plug-ins

Aurelien's Image J plug-ins

1. Moving average (simialr to the running average macro by Jens)

2. Thresholding Fluorescence image stacks.
  • a) Threshold determined by mean intensiity and standard deviation of the whole image. the threshold value is determined by n*s.d., where n is an user-input value.
  • b) same principle but the threshold value is locally determined. A frame is divided by grids and local threshold is determined in each position.
3. Radial-kyomograph: plots radial intensity profile (x) vs Time of a sequence.

4. Intensity profile dynamics in a segment of radius (used for FRAP or FLIP).

5. 3D tracking, center of mass search in xy plane + max intensity serach in z axis.

Wednesday, March 22, 2006

Tracking based on Particle analyzer in ImageJ

Algorithm: "tracking2D_particleAnalV1.txt"

depending on the precious positions, seraching origin point is estimated. starting with the original ROI size (currently fixed to 13x13. optimized for the size of Gaspar's embryo), particle analysis function is applied. Objects touching the edge of the ROI is excluded for this searching. if the ROI perfectly contains the thresholded embryo, then the particle is recognized and the centroid is calculated.
A. analyze particel within ROI (exclude touching edge)
A-1 if there is one particle (nResults==1), then the centroid is recorded.

A-2 if there is no particle (nResults==0), then
A-2-1. Do "particle anazysis"without "exclude touchin edge"
test wether the object is there, but overlapping with the ROI edge.

A-2-1-1. if this analysis returns one particle, then the particle is overlapping with the ROI edge.

--> increase the ROI size by 1 pixel and go back to A.

A-2-1-2. if this analysis returns no particle

A-2-1-2-1. the threshold level is too high, the segmented object does not have a sufficient size.
-- test changing the threshold to a lower level, then do particle analysis. Iterate this until threshold==1
A-2-1-2-1-1. >1 particel found at a certain threshold.
--> reset the threshold to the discoverd lower value, and go back to A.

A-2-1-2-1-2. if there is no particle even for a threshold level ==1, then There is no particle.

--> revert the threshold to the original level, increase the ROI size by 1 pixel
and then

A-2-1-2-1-2-1. if the ROI size is still within the scan range, then the particle is outside the ROI.
--> increase the ROI size by an increment, go back to A.
A-2-1-2-1-2-2. the object went out of the frame (or too far away).
--> terminate the tracking, suggest consideirng a larger scan range for a possible high speed.

A-3 if there are multiple particles
-- test changing the threshold to a lower level, then do particle analysis. Iterate this until threshold==1
A-3-1 if single particle could be segmented (nResults==1and the area is not too big),
threshold is too low, so the single particle is segmented as multiple particles.
--> reset the threshold to the discoverd lower value, revert the ROI size to the minimum and go back to A.

A-3-2 if multiple particles (nResutls>1 or area too big),
--> there are actually several particles.

A-3-2-1 (if THIS action is not done previously)
shift the position of ROI to the extended direction of estimated movement,
revert the threshold level, revert the ROI to the minimum size and go back to A
A-3-2-2 if A-3-2-1 is done once already, its impossible to segment one single particle
-- Terminate the tracking. (suggest particles are crossing).
POSSIBLE solution. increase the threshold level, revert the ROI size and go back to A.

B Estimate the next pposition, and starting from that position in the next frame, go back to A

Tuesday, March 21, 2006

Jerome, Jacomine


multiple particle tracking program in IgorPro. An approach similar to Danuser's specle fluoresscence microscopy.

screening of different types of signal distribution against nucleus - dispersed or concentrated. --> better with an estimation of the distance.--> Olmpus ScanRanlysis. (talk to Rainer)

Wednesday, March 08, 2006


Published online: 5 March 2006; Nature Cell Biology
Coordination of microtubule and microfilament dynamics by Drosophila Rho1, Spire and Cappuccino
Alicia E. Rosales-Nieves1, 2, James E. Johndrow1, 2, Lani C. Keller1, 2, Craig R. Magie1, Delia M. Pinto-Santini1 & Susan M. Parkhurst1

"Here, we demonstrate that Capu and Spire have microtubule and microfilament crosslinking activity."


Monday, March 06, 2006

Comparison of Quicktime (mov) compressions::

Adobe Premier was used for testing

MPEG-4 video or MotionJPEG-B at higher quality, or sorrento3.
File size-wise, MPEG4 is the smallest (e.g. 770k), then Sorrento3 (963k) MotionJPG-B (1231k) is the largest.

sorento (2)
- limit data rate to 100kb/sec
- "quality" does not really matter, 30% is almost equal to 100%
--> small structures disappear.
--> contrast becomes high. 10 point texts cannot be resolved.

limit data rate to 100kb/sec
"quality" does not really matter, 30% is almost equal to 100%
---> best quality

MPEG-4 video
-- very poor quality, with "quality" at 30%. at 70%, OK.
-- file size is incedibly small.

MJPG A or B (motion JPEG)
-- vertical lines appear after compresion with 30% quality.
-- with 70% quality, small structures are resolveable.File size becomes more than two times the 30% file.

video codec

I looked for ways to produce .mpg movies and follwoing is a list of related URL. I decided not to make .mpg movies, but informaiton here might be useful for those who want to do similar things.

* Free, A great site to look for video encoders/decoders.

* Detailed Explanations on different codecs

* A List of Codecs is at Very useful

*Stinky's MPEG-2 Codec

AviDemux 2.1.1 (free, but windows version seems to be not working well)
"AviDemux is a video editor and encoder which can edit, encode, requantize MPEG and AVI, including DivX. It is almost like VirtualDub, but can also encode to VCD/SVCD/DVD mpg."

MPEG Mediator 1.5
(did not try)
"MPEG Mediator supports MPEG-1 and MPEG-2 streams (Transport Streams are not supported) including AC-3, MPEG and LPCM audio-substreams. MPEG Mediator provides a great amount of flexibility through its use of 'Flask'-style plugins to output the converted video ('Flask' is popular MPEG stream converter)."

YMPEG 3.1a (limited to a 3hrs trial)
YMPEG is an award winning professional codec which can integrate itself with Windows and offers seamless encoding from your favorite application (VFW). This powerful and unique feature of YMPEG makes it possible to encode video from virtually any video applications like VirtualDub, Adobe Premiere, AE, Vegas, DVDx, K!TV etc. (using Windows Compression Manager).

WinMPG Video Convert 6.2.1 (unregistered version overlays texts in the output)

Windows Media Encoder 9.0
--> outputs WMV formats

Focus Video Converter 2.1
--> saying ts free in "" but actually you have to buy for full seatures.
--> only mpeg1 and 2.

--> Conversion is fast.The output is mpeg. Outputs with fixed aspect ratios for NTSC, PAL or film.

Friday, March 03, 2006

Snow -> Rain in Heidelberg

Since morning it was heavily snowing but then from 14:00, snow turned into rain. If road surface will be frozen, famous disaster down the Steigerweg is expected (crashed cars, abandoned cars..)

(1) Bleaching Correction macro
A year ago or so, Jens Rietfdorf (now in Basel) made a macro to correct for Bleaching of fluorescence of image series. Julien Colombelli asked me for different reason, some darkening of the image due to whatever camera setting or so, but the same macro should be working for this and also for all other similar decay of signal. This imageJ macro is downloadable from EAMNET website.

(2) FRAP
Fatima Verissimo asked me about FRAP fitting. Before doing experiment, go through chapter 2 in FRAP maual to see what kind of data (what you should measure out of FRAP image sequences) is required. I discussed with fatima about how to do FRAP analysis when the bleached ROI moves around after bleaching. This is a challenging task. The best way would be to label the target structure with a probe with a different fl. wavelength from the probe one is using for the FRAP. This none-FRAPped fl. signal would then be used for tracking the position of the structure. If the second probe is not possible... then that will be something to think about how to do.

(3) Going through Heiko's paper on Cholesterol. Great data, great paper. The project is also techincally rich FRAP, fl. kinetic studies, Modeling etc etc.

(4) 2D cross-correlation fixing. Somthing is wrong, probably small mistakes in back calculating 1D array to 2D arary (xy coordinates).

Wednesday, March 01, 2006

new papers

cortical flow related (Bray):

Nature Cell Biology 8, 216 - 226 (2006)
Published online: 26 February 2006; | doi:10.1038/ncb1367

Myosin II functions in actin-bundle turnover in neuronal growth cones
Nelson A. Medeiros, Dylan T. Burnette & Paul Forscher

Nature Cell Biology 8, 264 - 270 (2006)
Published online: 12 February 2006; | doi:10.1038/ncb1370

A microtubule-binding domain in dynactin increases dynein processivity by skating along microtubules

Tara L. Culver–Hanlon1, Stephanie A. Lex1, Andrew D. Stephens1, Nicholas J. Quintyne2, 3 &
Stephen J. King1

FRAP modeling::
Biophysical Journal 90:1878-1894 (2006)

Dissecting the Contribution of Diffusion and Interactions to the Mobility of Nuclear Proteins

Joël Beaudouin, Felipe Mora-Bermúdez, Thorsten Klee, Nathalie Daigle and Jan Ellenberg

Biophys. J. 2006 90: 2213-2220

Wei-hui Guo, Margo T. Frey, Nancy A. Burnham, and Yu-li Wang
Substrate Rigidity Regulates the Formation and Maintenance of Tissues