Need help on translocation assay


#1

Hi all,
First of all, I want to say a big thanks to you and your entire team who have developed this awesome software. And to those who respond to forum topics with really good advice.

The advices, the forum, the tutorials has really given me the basic ground to use CP. And I’ve only been using CP for less than 15 days.

I need some help with my first run. I’m working on a cytoplasm-nucleus translocation assay and presently, using the example given on the website.
I successfully run the pipeline on the example set of SBS images. And got a output called, “testout”.
When I do exportdata from DataTools menu, it asked, which measurements to export.
My question is, what are these measurements namely image, nuclei and all?
Which one are most useful to examine for the assay? And are they also known as ‘object type’?

Also, it will be immense help to me, if there is any explanation for the the variables given in the output file per say, image object.

waiting eagerly for your reply


Which object analysis to choose for translocation assay
#2

Hi,

Each of the measurements listed in the ExportData dialog box is based an object defined and generated by the pipeline. For example, Nuclei is created by IdentifyPrimAutomatic (module #6). Have a look at the individual modules to see where the other objects come from.

For each object, all the measurements associated with that object are exported to the Excel file (or whatever format you choose). For example, for the Nuclei object, you have the following:

  • Object number, center locations: Generated by IdentifyPrimAutomatic (module #6)
  • Secondary object number: Generated byIdentifySecondary
  • Correlation measurements: Generated by MeasureCorrelation, between CorrGreen and CorrBlue
  • Intensity measurements: Generated by MeasureObjectIntensity, one for CorrGreen, one for CorrBlue
  • Morphological measurements: Generated by MeasureObjectAreaShape
  • Texture measurements: Generated by MeasureTexture, one for CorrGreen, one for CorrBlue

For the exported file for the Image object, these measures are taken as a mean of the individual object measurements.

For a translocation assay, you are mostly likely interested in the relative amounts of protein distributed between two compartments, such as the nucleus and cytoplasm (I would suggest looking at our paper in GenomeBiology (genomebiology.com/2006/7/10/R100) for an good example). So you can look at the fraction of the GFP fluorescence intensity represented in the nucleus vs. the cytoplasm, i.e, the results of the CalculateRatios module (named Ratio1) which would be in the exported output for the Nuclei object.

Also, the module ExportToExcel at the end of the pipeline will give you the same result as using ExportData on the output .mat file.

Hope this helps,
-Mark


#3

Thanks Mark,
It was a real help to me.
I was already looking into all those figures in my Excel output.
your description is really helping me out to do them.

I’ll ask more wherever I got stuck-up.

But can u tell me,

  1. how to assay the quality of the experiment since there V’ and Z values to figure that out.
    In the excel file of my Experiment object, there are V, Z, EC50 for each of the origThreshold, Threshold, ObjectCount, Classify, Children, Correlation, Intensity, Areashape and Texture.

Can u briefly give a one-liner for each of these.

  1. how to quantity the percentage of cells translocated to nucleus, depending upon Ratio1?

  2. Also, whats channel1ILLUM.mat and channel2ILLUM.mat. I dont have these files with me to load. How should I work with my ‘bitmap’ images for two channels: red & green.

I Think, explanation of these will give me pretty gud idead to handle all such measurements for other experiments too.

Thanks & regards
Mridul KK


#4

Hi,

After running the pipeline given for SBS images, I exported excel files for each object: Images, Nuclei, DistCytoplasm, DistanceCells, PropCells, PropCytoplasm, Thresholded cells and Experiment.

From the Image-based excel file, I generated a graph (attached below).
Can you tell me, what exactly, I can interpret from it in order to see, how much translocation has occurred in response to SBS doses.

I want to be confirm for such interpretations before moving forward for my own experiments.
Thanks & waiting for your reply.

Can you also reply to my previous post (Fri Jun 27, 2008 1:56 pm).
regards
Mridul KK



#5

Hi ,

I’ll take each of these in order:

(1) Assay experiment quality: The CalculateStatistics module generates the V and Z’ factors for all measurements made by the pipeline. But not all these factors are relevant for a translocation assay. (a) OrigThreshold and Threshold are used to segment the cells from the background (b) ObjectCount are the number of objects found in a given image. © Classify comes from the ClassifyObjects module which separates measurements into bins based on a metric (d) Children comes from the IdentifySecondary module which produces the cell boundaries. (d) Correlation, Intensity, AreaShape and Texture come from the Calculate modules of those names. For a translocation assay, you will probably be interested in (d).

(2) Percentage of cells with nuclear translocation: Ratio1 gives you the blue-to-green florescence proportion on a object-by-object basis in the Excel file. The threshold for what ratio value counts as sufficient translocation for a particular cell is up to your judgment. Once you decide this, you can count cells above and below this threshold for a given dose, and come up with a curve much like Figure 4b in the Genome Biology paper I mentioned previously.

(3) channel1ILLUM.mat and channel2ILLUM.mat. These are illumination correction image files. Since in microscopy, there is commonly uneven background illumination. If it is not accounted for and corrected, it may distort the fluorescence values leading to incorrect quantification.

The modules CorrectIlluminationCalculate and CorrectIlluminationApply deal with this, and are incorporated in the additional file ExampleSBSIlluminationPIPE.mat that should be in the same directory as the SBS images. The result of this pipeline are the ILLUM.mat files which are then used by ExampleSBSPIPE.mat. Since you ran the ExampleSBSPIPE pipeline successfully, I assume that you found them (or at least the pipeline did)? If you don’t have them, you can get them with the rest of the files: cellprofiler.org/linked_files/Ex … Images.zip

You will need to create illumination files for your particular red and green images. Have a look at the ExampleSBSIlluminationPIPE pipeline to see how it works, and how the two channels are fed in.

(4) Interpretation of graph. The amount of the SBS dose is given by the ‘SBS dose’ bar. The amount of translocation increases with increasing SBS dose. The way you can see this is looking at the correlation between the nuclear fluorescence in the blue and green channels, as shown by the red bar.

For the low does, protein is exported out of the nuclei so the nuclei are bright in the nuclei channel (Channel1: blue) but dark in the GFP channel (Channel2: green). Hence the fluorescence in CorrBlue and CorrGreen are anti-correlated and is negative. At the higher doses, export is blocked and protein accumulates in the nucleus, so the nuclei are bright in both channels. Hence, CorrBlue and CorrGreen are positively correlated.

Note that this graph comes from the Image Excel file, which means that it is averaged over all nuclei in that image. If you want to see the individual nuclei data, you need to see the Excel file for Nuclei (as mentioned for #2).

Hope this helps,
-Mark


#6

Thanks Mark,
It was a great help and I’m so happy for all yr support.
I was away from lab for past one week, so cudn’t reply.

I’ll go through yr suggestions and will get back to you, if any query.
Thanks once again
Mridul KK


#7

An additional note about (2): using the ClassifyObjects module, you can have CellProfiler automatically count cells as ‘positive’ or ‘negative’ based on the threshold you have chosen for the feature you have chosen (the ratio) and then calculate the % positive.
Anne