Help - Analysis of Large Number of Images

[KeithEPrater]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Dear list,

I've recently become involved in a project that is going to require a large
number of images to be analyzed for multiple antigens. The images will be
section series (approx. 6 - 12 um deep) of porcine tissue captured on a
Leica confocal system, indirectly stained using 3 different fluorophores. I
will need to retrieve volume measurements from the section series. From the
number of tissues I will receive, I will be acquiring hundreds of images.

My usual method for smaller workloads has been to load the raw .tiff files
into Nikon Elements, and go through the time-consuming process of telling
the software how to handle the series (wavelengths, steps, etc.), manually
calibrate the image from the Leica text output, apply median filter,
threshold and run measurements. I have as many of those steps as possible
rolled into one macro. Also, Elements has the ability to automatically
reconstruct the section series based on file name, but this is unreliable
with files captured using a sequential scan.

Does anyone out there have experience with a similar situation and found a
method that is more efficient for dealing with such a large number of
images? Is there a better software alternative? I think the biggest
time-hog is having to take raw image files from a Leica system and
reconstruct them in Nikon Elements. Perhaps Leica has an analysis package
better suited for this?

Just wanted to see if anyone else has fought this battle.

Thank you all in advance.

Keith Prater
Research Technician II
Center for Environmental Biotechnology
University of Tennessee-Knoxville

[Justin Price]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Keith,

You should take a look at ImageJ or Fiji.  It sounds like you have a grasp on the basics of image processing, and you have an idea of what you are looking for so I would give it a shot.  Here are a few advantages to the ImageJ/Fiji approach.

-freely available

-bioreader has many proprietary formats available to read raw data as well as metadata

-easy to get started writing macros with the "record macro" function

-access to hundreds of niche programs and macros that could cost thousands of dollars is purchased via commercial route

Disadvantage to ImageJ/Fiji:

-some programs may not work as expected or at all depending on how you files are labeled or formatted

-customer support is on a voluntary basis so find a friend to help bail you out if you get stuck.

FIJI has more 3d applications for making the measurements you are looking to do included with the initial download than does ImageJ so I would check this out first.

Enjoy!
Justin  

On Mar 14, 2013, at 10:15 AM, Keith Prater <[hidden email]> wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Dear list,
>
> I've recently become involved in a project that is going to require a large
> number of images to be analyzed for multiple antigens. The images will be
> section series (approx. 6 - 12 um deep) of porcine tissue captured on a
> Leica confocal system, indirectly stained using 3 different fluorophores. I
> will need to retrieve volume measurements from the section series. From the
> number of tissues I will receive, I will be acquiring hundreds of images.
>
> My usual method for smaller workloads has been to load the raw .tiff files
> into Nikon Elements, and go through the time-consuming process of telling
> the software how to handle the series (wavelengths, steps, etc.), manually
> calibrate the image from the Leica text output, apply median filter,
> threshold and run measurements. I have as many of those steps as possible
> rolled into one macro. Also, Elements has the ability to automatically
> reconstruct the section series based on file name, but this is unreliable
> with files captured using a sequential scan.
>
> Does anyone out there have experience with a similar situation and found a
> method that is more efficient for dealing with such a large number of
> images? Is there a better software alternative? I think the biggest
> time-hog is having to take raw image files from a Leica system and
> reconstruct them in Nikon Elements. Perhaps Leica has an analysis package
> better suited for this?
>
> Just wanted to see if anyone else has fought this battle.
>
> Thank you all in advance.
>
> Keith Prater
> Research Technician II
> Center for Environmental Biotechnology
> University of Tennessee-Knoxville

[mahogny]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

with the obvious bias of being one of the authors, I can just point you
toward
http://www.endrov.net
and the new high-throughput analysis tool.

I strongly recommend storing the images already at the source into a decent
file format, which better keeps track of the relationship between your
images. Only if you are at loss with this should you do a name based import
(why is it unreliable with your input images? if the naming is systematic,
it should not be). Maybe you also want to consider storing in OMERO while
you are at it, as a cleanup phase after the acquisition

/Johan



On Thu, Mar 14, 2013 at 3:15 PM, Keith Prater <[hidden email]>wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Dear list,
>
> I've recently become involved in a project that is going to require a large
> number of images to be analyzed for multiple antigens. The images will be
> section series (approx. 6 - 12 um deep) of porcine tissue captured on a
> Leica confocal system, indirectly stained using 3 different fluorophores. I
> will need to retrieve volume measurements from the section series. From the
> number of tissues I will receive, I will be acquiring hundreds of images.
>
> My usual method for smaller workloads has been to load the raw .tiff files
> into Nikon Elements, and go through the time-consuming process of telling
> the software how to handle the series (wavelengths, steps, etc.), manually
> calibrate the image from the Leica text output, apply median filter,
> threshold and run measurements. I have as many of those steps as possible
> rolled into one macro. Also, Elements has the ability to automatically
> reconstruct the section series based on file name, but this is unreliable
> with files captured using a sequential scan.
>
> Does anyone out there have experience with a similar situation and found a
> method that is more efficient for dealing with such a large number of
> images? Is there a better software alternative? I think the biggest
> time-hog is having to take raw image files from a Leica system and
> reconstruct them in Nikon Elements. Perhaps Leica has an analysis package
> better suited for this?
>
> Just wanted to see if anyone else has fought this battle.
>
> Thank you all in advance.
>
> Keith Prater
> Research Technician II
> Center for Environmental Biotechnology
> University of Tennessee-Knoxville
>

--
--
-----------------------------------------------------------
Johan Henriksson, PhD
Karolinska Institutet
Ecobima AB - Custom solutions for life sciences
http://www.ecobima.com  http://mahogny.areta.org  http://www.endrov.net

<http://www.endrov.net>

[Johannes Schindelin]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Justin,

On Thu, 14 Mar 2013, Justin Price wrote:

> Disadvantage to ImageJ/Fiji:
>
> [...]
>
> -customer support is on a voluntary basis so find a friend to help bail
> you out if you get stuck.

If you give [hidden email] a chance to help you, I am sure you will
be delighted at the helpful feedback you will get there.

Ciao,
Johannes

[KeithEPrater]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi folks!

Thanks for the kind replies and suggestions.

I'm leaning towards the ImageJ/Fiji route, due to budget constraints. I
really appreciate all of the advice and will keep other suggestions in my
notes.

Cheers!

On Thu, Mar 14, 2013 at 12:53 PM, Johannes Schindelin
<[hidden email]>wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi Justin,
>
> On Thu, 14 Mar 2013, Justin Price wrote:
>
> > Disadvantage to ImageJ/Fiji:
> >
> > [...]
> >
> > -customer support is on a voluntary basis so find a friend to help bail
> > you out if you get stuck.
>
> If you give [hidden email] a chance to help you, I am sure you will
> be delighted at the helpful feedback you will get there.
>
> Ciao,
> Johannes
>

[Schebique]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi.
I think that analysis should be able to automatize in nis elements as well.
Feel free to contact guys from laboratory imaging company. I am sure they
will be very open to help you with automatization.

However imageJ based analysis is of course viable solution :)

Ondrej

[mahogny]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

> > Disadvantage to ImageJ/Fiji:
> >
> > -customer support is on a voluntary basis so find a friend to help bail
> > you out if you get stuck.
>
> If you give [hidden email] a chance to help you, I am sure you will
> be delighted at the helpful feedback you will get there.
>


I think this is a misperception we need to work on.

* If you buy software from company X, the only company that can really work
on your issues (including fixing bugs) is the company X

* If you use open source software, you can get help from the community. And
you can get help from company 1,2,3,4,5,6,7,8,9 and 10. None of these made
the software, but since the software is open source, any company has the
opportunity to fix any bug or implement any feature you want. This is a
level of support you simply cannot expect from any large company, unless
the feature you need will become their next big selling point.

So instead of getting stuck with company X, and be forced to pay for a
software that in worst case cannot do what you need, maybe rather go with
the free and flexible alternative, and only if you have to, pay later?

/Johan
(with the bias of being one of the many available consultants selling
commercial support for open source software)

--
--
-----------------------------------------------------------
Johan Henriksson, PhD
Karolinska Institutet
Ecobima AB - Custom solutions for life sciences
http://www.ecobima.com  http://mahogny.areta.org  http://www.endrov.net

<http://www.endrov.net>

[Rosemary.White]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Dear all,

Have any of you used aptamers (up to 100-nucleotide polymers of either RNA
or DNA) instead of antibodies in localisation work? There are a couple of
published papers, but seems that they've mostly been used on gels or in
pull-down assays and the like. Could be good but would be interested in
any feedback. Expensive to develop at first, but then possibly much
cheaper to re-order. I've ordered a couple of published ones to try.

cheers,
Roseamry

Dr Rosemary White
CSIRO Plant Industry
GPO Box 1600
Canberra, ACT 2601
Australia

T 61 2 6246 5475
F 61 2 6246 5334
E [hidden email]


p.s. sorry for cross-post with microscopy list.

[George McNamara]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Rosemary,

I have not used aptamers - yet. The lab I am joining at MDACC in Houston
has a collaboration with Base Pair Technologies, also in Houston. Base
Pair makes DNA aptamers. I anticipate doing aptamer fluorescence
detection with them later this year. Their white paper
http://www.basepairbio.com/wp-content/uploads/2012/04/WhitePaper_BasePairBio.pdf
Immunohistochemistry section cites two papers:

    "Employing aptamers in conventional histology and molecular imaging
    is another logical extension for their use in lieu of antibodies. In
    recent years, directly, or indirectly labeled aptamers have, for
    instance, been used to detect individual receptors [77] or to
    differentiate cancerous from non-cancerous cells [78]."
    77. Chen Y, Munteanu AC, Huang Y-F, Phillips J, Zhu Z, Mavros M, Tan
    W: Mapping receptor density on live cells by using fluorescence
    correlation spectroscopy. Chemistry 2009, 15:5327-5336.
    78. Chen HW, Medley CD, Sefah K, Shangguan D, Tang Z, Meng L, Smith
    JE, Tan W: Molecular recognition of small-cell lung cancer cells
    using aptamers. ChemMedChem 2008, 3:991-1001.

Their "Aptamer blots" section mentions conjugating their aptamers to
quantum dots or biotin. For the latter, then streptavidin-HRP detection.
Speaking of HRP, I plan to test "HRP80" in the next few months, either
with Base Pair's aptamers, or antibodies. HRP80 is from Fitzgerald,

    http://www.fitzgerald-fii.com/streptavidin-poly-hrp80-conjugate-65r-s118.html
    Streptavidin Poly-HRP80 Conjugate
    Catalog No. 65R-S118
    Streptavidin Poly-HRP Conjugate is streptavidin biotin-binding
    protein that is conjugated with polymers of horseradish peroxidase,
    enabling signal amplification and detection of biotinylated
    antibodies for IHC and other methods. This poly-HRP conjugate is
    designed to deliver the highest sensitivity and low background in
    immunoassays where sample volume is limited or when the target
    molecule is present at low levels. The estimated average number of
    HRP monomer molecules in SA-PolyHRP20 conjugate is 100 (20 X 5), in
    SA-PolyHRP40 - 200 (40 X 5) and in SA-PolyHRP80 - 400 (80 X 5).
    Thus, PolyHRP brings in reaction with substrate development system
    much larger number of enzyme label molecules (per one bound analyte
    molecule) than conventional conjugates do.

I am a big fan of fluorescent tyramide signal amplification, so the
prospects of ~400 HRP's per streptavidin is appealing, at least for cell
surface applications. My expectation is that this should provide single
antigen molecule detection. There are 5 tyrosines per HRP (though not
all on the surface), so the HRP polymer component alone could provide up
to 2,000 covalent binding sites for the tyramide. As long as there is no
non-specific binding, should be great.

**

Somalogic has published several papers on aptamer-based fluorescence
microscopy detection of specific antigens. You can check their web site
for references. Their specialty is to include modified bases in their
aptamers, which make it more expensive.

**

Samie Jaffry has 2 Science papers (Paige et al 2011, 2012) on "Spinach"
RNA aptamer for live cell imaging. Uses the same modified tripeptide as
EGFP (commercially available peptide).

Best wishes,

George





On 3/17/2013 12:57 AM, [hidden email] wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Dear all,
>
> Have any of you used aptamers (up to 100-nucleotide polymers of either RNA
> or DNA) instead of antibodies in localisation work? There are a couple of
> published papers, but seems that they've mostly been used on gels or in
> pull-down assays and the like. Could be good but would be interested in
> any feedback. Expensive to develop at first, but then possibly much
> cheaper to re-order. I've ordered a couple of published ones to try.
>
> cheers,
> Roseamry
>
> Dr Rosemary White
> CSIRO Plant Industry
> GPO Box 1600
> Canberra, ACT 2601
> Australia
>
> T 61 2 6246 5475
> F 61 2 6246 5334
> E [hidden email]
>
>
> p.s. sorry for cross-post with microscopy list.
>
>