Posted by
George McNamara on
URL: http://confocal-microscopy-list.275.s1.nabble.com/DIC-or-Phase-contrast-for-image-morphometry-mesasurements-tp7068286p7073031.html
*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy*****
Hi Monique,
You're welcome. I see that you are interested in shape factors. If you
have access to MetaMorph software, the ones you mentioned are in there
(plus fiber length and breadth, though yeast are probably not snake like
enough to make these useful) and 40+ of my favorites: Elliptical Fourier
Analysis (EFA). Page 7 of
http://sydney.edu.au/medicine/bosch/facilities/microscopy/ImagingSoftware/IMA_Measurement_Parameters.pdf
is not a complete list of MetaMorph's EFA parameters (I take no credit
or blame for any of the drawings).
There is an option to turn them on in Preferences. You can also combine
interesting coefficients, as Tom Coates did in a 1993 paper (before I
went to work for him) - PubMed 8243213. See also the classic paper by
Ferson, Rohlf et al 1985 -
http://www.jstor.org/pss/2413345 (I turned
Rohlf's Fortran code into Turbo Pascal, which the MetaMorph programmers
turned into long forgotten MPL code, which morphed into something that
still works in MM7.7). Rohlf's web site is still up, and even has 3D EFA
-
http://life.bio.sunysb.edu/morph/soft-outlines.html (I only
implemented 2D - also Rohlf's code and mine were able to average
harmonics over all inputted perimeters and then do the inverse FFT to
make an averaged shape - never implemented in MetaMorph ... when I
learned of TypIC - see below - I liked the approach of picking the
median, since that would be a real shape (if odd number of cells
inputted)).
Robert Murphy's PSLID and related software/database at
http://murphylab.web.cmu.edu/ may or may not have EFA built in to it. I
don't see whether PSLID or SLIC still allows uploading ~100 images and
calculate results for you. The ancestor of PSLID was TypIC:
Toward objective selection of representative microscope images.
</pubmed/10096918> Markey MK, *Boland* MV, *Murphy* RF. Biophys J. 1999
Apr;76(4):2230-7. PMID: 10096918.
Scientists wishing to communicate the essential characteristics of a
pattern (such as an immunofluorescence distribution) currently must make
a subjective choice of one or two images to publish. We therefore
developed methods for objectively choosing a typical image from a set,
with emphasis on images from cell biology. The methods involve
calculation of numerical features to describe each image, calculation of
similarity between images as a distance in feature space, and ranking of
images by distance from the center of the feature distribution. Two
types of features were explored, image texture measures and Zernike
polynomial moments, and various distance measures were utilized.
Criteria for evaluating methods for assigning typicality were proposed
and applied to sets of images containing more than one pattern. The
results indicate the importance of using distance measures that are
insensitive to the presence of outliers. For collections of images of
the distributions of a lysosomal protein, a Golgi protein, and nuclear
DNA, the images chosen as most typical were in good agreement with the
conventional understanding of organelle morphologies. The methods
described here have been implemented in a web server
(
http://murphylab.web.cmu.edu/services/TyplC).
{web link long dead}.
Best wishes,
George
p.s. if you can find a MetaMorph CD, there should be an EFA bibliography
on it - probably inside a zip file. I can send it to anyone interested
(may be a few days due to travel).
*****
Hi Mike,
We are trying to caracterize and classify different yeast mutants by their morphology appearance (morphotype) and size. Actually, we are trying to caracterize in 2D with the different shape factors like circularity, elongation, aspect ratio, sphericity... and so on, and by the area. For your volume measurements do you need 3D acquisition? I guess volume should be more precise than area. Yes, your publication interests me. Thanks a lot
All the best,
Monique Vasseur
On 12/7/2011 10:13 AM, Vasseur Monique wrote:
> Hi George,
>
> Thanks so much for the advices. I didn't think about blue monochromatic illumination to improve resolution neither to use a test plate. You comments are always much appreciated and so useful. Thank you!
> Best wishes,
>
> Monique
> -----Message d'origine-----
> De : Confocal Microscopy List [mailto:
[hidden email]] De la part de George McNamara
> Envoyé : 6 décembre 2011 21:45
> À :
[hidden email]
> Objet : Re: DIC or Phase contrast for image morphometry mesasurements?
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
>
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy> *****
>
> Hi Monique,
>
> Brightfield with optimized Kohler illumination (condenser field aperture
> diaphragm focused and centered, numerical aperture diaphragm optimized
> for the NA of your objective lens). You did not specify what detector -
> I will assume a good scientific grade monochrome digital camera (ex.
> Hamamatsu ORCA-ER). Use monochromatic light (ex. 546 nm narrow bandpass
> filter), or the emission filter of a green or blue fluorescence bandpass
> filter cube (ex. GFP or DAPI). Shorter wavelength is better. Higher NA
> objective AND condenser lenses better. Ideally use a 1.4 NA oil
> immersion objective lens and oil immersion objective lens. I recommend X
> and Y pixel size 3.5x smaller than the resolution equation of d =
> 1.22*lambda/(NAobj+NAcond), that is pixel size ~ d/3.5 (note; if you
> cannot get to 3.5x, try to get close).
>
> Refractive index of the mounting medium (culture medium if live cells)
> is tricky. Higher NA, closer to that of the lens immersion medium is
> better, but:
> 1. may be higher osmolarity, resulting in cell shrinkage, which may
> defeat the goal of making the measurement.
> 2. the cells will become closer to invisible, the closer to R.I. matched
> and best focus you reach (reference is F. Zernike, Science 121 (Issue
> 314, Mar. 11, 1955: 345-349, http ://www.jstor.org).
>
> With respect to #2, a good digital camera and understanding of
> acquisition settings, contrast adjustments, and shading correction, will
> get you excellent and reproducible images, even if contrast by eye is
> minimal. Some DAPI or Hoechst (or GFP positive cells) may help find the
> cells!
>
> If you have access to a Diatom 'test plate' - Carolina Biological Supply
> (good luck with CBS's search engine) - see
>
>
> Fundamentals of light microscopy and electronic imaging
>
> By Douglas B. Murphy
> pages 93-94
>
>
http://books.google.com/books?id=UFgdjxTULJMC&pg=PA93&lpg=PA93&dq=diatom+test+plate&source=bl&ots=vkHltkLVvi&sig=L399EbcUIx3ebJac54Hfvce5vxM&hl=en&ei=887eTsjnNpTAtgeT0cmFBA&sa=X&oi=book_result&ct=result&resnum=2&ved=0CCQQ6AEwAQ#v=onepage&q=diatom%20test%20plate&f=false>
>
> g and h specimens(h is Amphipleura pellucida, 0.27 um/stria) are hard to
> resolve with white light, better with green (GFP emission filter),
> better with blue light (DAPI bandpass emission filter ... typically
> halogen lamp should be at max voltage and will likely need to optimize
> the exposure time).
>
> This slide as well as a thin unstained muscle tissue section slide were
> great slides to work with at UIC (video contrast adjustment days). For a
> while Richardson (Corp?) in Canada had a nice resolution target. Other
> companies still offer them. Measuring the performance on the diatom
> slide or other test target and stating that measurement in your methods
> section should help the review process when you go to publish.
>
>
> Plan R: reflection mode confocal microscopy with a short wavelength
> laser line (405 nm or more practically on most Argon ion laser systems,
> 458 nm). Be careful to record the PMT gain and offset settings. Note
> that if a cell is very close to the coverglass, (or slide) you can get
> interference reflection contrast (helpful or not depending on goals).
> IRM is useful to find the coverglass, but may cause confusion for cells
> on or near the coverglass.
>
> Plan S: plasma membrane (or cell wall) fluorescent label using a
> fluorophore or fluorescent protein that you can use on your nearest STED
> nanoscope. Alternatively, PALm, STORM, FPALM, 3D-SIM, etc.
>
> Best wishes,
>
> George
>
>
> On 12/6/2011 3:23 PM, Vasseur Monique wrote:
>
>> *****
>> To join, leave or search the confocal microscopy listserv, go to:
>>
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy>> *****
>>
>> Dear All,
>>
>> I wonder what is better for yeast morphometry shape and size image analysis:
>> should we better use phase contrast images or DIC images? Is one more
>> accurate for measures? )Which threshold will be more precise: the one of DIC
>> shear or the one of phase contrast halo?) Do some of you have experience on
>> this question? Any input is welcome Thanks a lot!
>> Monique
>>
>>
>>
>
>
--
George McNamara, PhD
Analytical Imaging Core Facility
University of Miami