Help for analysing many sites for each well of a 96 plateplt


#1

I have two channels and am scanning six sites in each well of a 96 well plate. My goal is to measure nuclear cytoplasmic translocation with a drug screen of 122 drugs, done in triplicates. I guess the best way to do so is analyse plate by plate.
I am unable to setup a metadata profile for each plate to pull metadata from the file name and also where I can tell profiler to put the sites together in well for each channel. Can you help me with an example or a pipeline


#2

Hi,

Most likely all plates can be analyzed together if there is a set of unique specifier in the path and/or file names. Indeed you should be able to drag all your plates’ images in Images, and extract metadata using regular expressions in Metadata. Can you supply a couple images, or at least a few filenames so we can take a crack at writing the regexp, which is likely the holdup?

Thanks,
David


#3

Hi David,
I cant attach the files from my email as the file is too big to be attached here. The details of experiments are as follows: For the trial experiment, I used three wavelengths DAPI[w1], GFP[w2] and Cy3[w3] and scanned 4 sites for each well for each of the three wavelengths. The images are acquired with IXM Microscope and stored in tif formats. Normally I open the images in metaexpress software and save them from there to the powerpoint. For my upcoming big experiment, its not possible to manually do so for each site of each wavelength for each well. I plan to scan 8 wells for each wavelength for each well of a 96 well plate. then I will use CellProfiler to use the acquired images to do the analysis. The file names of images from one well[F04] is:
Plate1_F04_s2_w17292E940-8191-4EC8-8C6B-5F947DFFAA65.tif
Plate1_F04_s3_w1CD25C492-F511-4729-B304-338E063D4CD3.tif
Plate1_F04_s4_w101747325-83D2-478D-A18C-E74490A1C9BD.tif
Plate1_F04_s1_w1B3A88F5A-A83D-461F-897E-ABCD0B714E58.tif
Plate1_F04_s1_w3D93BAFDA-F64E-41A4-B737-CD5F9C63F30B.tif
Plate1_F04_s2_w3FB53C830-C6E6-40D7-A3E0-5A67D064A7E1.tif
Plate1_F04_s3_w3E656DE92-455C-4622-AC75-6B7F809A431E.tif
Plate1_F04_s4_w39C088E6F-F382-4EC2-AA36-868B92965855.tif
Plate1_F04_s1_w25CAE49B3-9ACB-431D-ACAC-7B2FB0900934.tif
Plate1_F04_s2_w270D1E04B-19A7-47C1-9F04-D253B0B90EEB.tif
Plate1_F04_s3_w2A1A31EB8-C671-4591-8B03-8B6974E69216.tif
Plate1_F04_s4_w2AAD239EC-4E4F-4878-BFF4-E3D40C0C154D.tif

I look forward to hearing from you.
Regards
Shivam


#4

Hi Shivam,

The attached pipeline sets up the Metadata module to what I expect will work for you. With it, you should indeed be able to run all plates of your data at once. Note though that this is best done with a computing cluster or a powerful workstation at least if you have more than a few plates.

Some notes:

  • The metadata filename matching should work with the filenames you provided.
  • The metadata folder name matching is a guess. The plate name needs to be extracted as well, and if it is MetaXpress then the plate name is likely the last folder name in the folder hierarchy. But you need to check this.

You can test both of the above by (1) Dragging a couple plates of data into the Images module (this can take awhile to load) (2) Click the Update buttons in Metadata and NamesAndTypes modules and then double-check that the output table columns matches the metadata you expect to extract for each metadata tag.

Let us know how it goes!
David
DL_metadata_setup.cppipe (3.32 KB)


#5

Thanks a lot David. Its a great help. I will let you know how it goes .
Bet Regards,
Shivam


#6

Hey,

This sounds similar to the image sets I am aiming to analyse. Still trying to get my head around the best way to analyse 100+ images from the one slide.

I have a pathway I am quite happy with for looking at nuclear/cytoplasmic translocation and measuring pixel intensity between nuclear and cytoplasmic regions.

I now have a major project which needs me to analyse at least 100 cells per 3 slide/staining conditions for many many patients. This means I have between 70-100 images per slide to combine which will become unwieldy if I have to run each image individually in cellprofiler.

DAPI Vectastain (BLUE “_B”)
FITC (Green “_G”)
Texas Red (Red “_R”)

I can’t attach the image type I get direct from the Zeiss Axion because it is in czi format.

Here is a .tif conversion of the 3-colour image:


I can’t attach the xml metadata file either that goes with the 3 colour tif :angry:

Tried as GIF also:


again I can’t attach the xml metadata file either that goes with the 3 colour gif :angry:

Cellprofiler AXION PBMC 060815.cpproj (941 KB)

I have been first opening Zeiss Axion images in the free Zen Lite software and converting the 3 colour image into 3 separate colour images identified into their colours with _B, _G and _R

Is there a way to convert three colour images into 3 tif images?

Do I even need to split up the three colour images before loading into cellprofiler pathway? I was under the impression that I had to do this and designed the attached pathway around this.

Split single colour images:




and a second image set:




Any advice or insight would be much appreciated.

Regards

Casey

University of Queensland
Australia


#7

I will have a try with the metadata cp pathway posted by David and see if I can prepare something. I suspect that it is beyond my very limited computer language skills.


#8

Hi Casey,

A few notes:

  • You can zip up any file type and the forum here will accept zip files

[quote]Is there a way to convert three colour images into 3 tif images?
[/quote]

ColorToGray module, using the ‘Split’ option

In general, yes. Most CellProfiler modules require grayscale inputs.

Does that answer your questions? Or did you have trouble extracting metadata like the posts above?
David


#9

Dear David,

Sorry for the delay in getting back to you. Thank you for your suggestions. I have a kernel of a plan forming of how I can do this.

I will give it a try and get back to you.

Regards

Casey

2x 3 colour files split into 3 (B, G, R indicate different colours):

030815_MND BIO 369_DAPI_P65_CD68_1_B.zip (182 KB)

030815_MND BIO 369_DAPI_P65_CD68_1_G.zip (191 KB)

030815_MND BIO 369_DAPI_P65_CD68_1_R.zip (199 KB)

030815_MND BIO 369_DAPI_P65_CD68_2_B.zip (167 KB)

030815_MND BIO 369_DAPI_P65_CD68_2_G.zip (195 KB)

030815_MND BIO 369_DAPI_P65_CD68_2_R.zip (201 KB)

4x 3 colour images:

Snap-1100_1.zip (2.56 MB)

Snap-1102_2.zip (2.56 MB)

Snap-1103_3.zip (2.74 MB)

Snap-1104_4.zip (2.6 MB)


#10

Back again after many days in the dark gathering images :imp:

Have made some further progress on the pathway thanks to your posted batch mode setup David. Thank you.

Getting stuck on the pathway when I run it in batch mode but not when I run it in single three set image mode.

Cellprofiler BATCH MODE FIRST ITERATION AXION PBMC 010915.cpproj (792 KB)

This is the error message I get:

Traceback (most recent call last):
File “cellprofiler\pipeline.pyc”, line 1934, in run_image_set
File “cellprofiler\modules\identifysecondaryobjects.pyc”, line 704, in run
File “cellprofiler\objects.pyc”, line 636, in add_objects
AssertionError: The object, GREEN_NUCLEI_AND_CYTOPLASM_CP, is already in the object set

Suspect it has to do with the naming of the files and/or the objects in the pathway but not sure why it only comes up when I try to run in batch mode.

Insight would be much appreciated.

Casey


#11

Hi Casey,

Just following up - was this answered by the other thread that you had open? cellprofiler.org/forum/viewtopic … 654#p13654

Thanks,
David