After the TMA-Deconvoluter has completed its operation, you will find yourself back at the main screen of the TMA-Deconvoluter (at the "Control") worksheet. You will notice that the output appears in additional worksheets within the TMA-Deconvoluter, and that a report of the activity of the TMA-Deconvoluter appears, as shown below:

Outputted Filenames

The "Processed" column indicates whether or not the raw scoring workbook was successfully processed. The "Output worksheet" column indicates the name of the worksheet within the TMA-Deconvoluter that contains the output data, and the "Output filenames" column contains the file names of the output files. If you had specified your own filenames, these would remain unchanged, and the corresponding files (4.txt, 5.txt in this example) would appear in the current working directory. If you had not specified any file names, the file names will be the same as the worksheet names.

Opening one of the output file names, or clicking on the corresponding worksheet tab within the TMA-Deconvoluter, should result in the screen below (if you had chosen to output in the PCL format):

PCL output file screenshot

Click on the image for a larger view of the picture.

For those of you familiar with the PCL (Pre-CLuster) format, you will recognize the formatting of the output.

• In Column A is the UID (Unique IDentifier) column, which contains the information passed on by TreeView to the Stainfinder program (for more information on this, please refer to the Stainfinder walkthrough).
• In Column B is the NAME column, which contains the information obtained from the lookup file. The information present is the same as Columns A-F in the lookup file, except with a pipe ("|") separating the different columns from that file. The unique case identifier, followed by the diagnostic information, appears in the field.
• In Column C is the GWEIGHT column, which defines the absolute weight each case is given in the clustering. The default value for each case is 1, indicating that each cases given equal weight in the hierarchical clustering. You may alter these values to some other number, prior to clustering.
• Columns D and onward: each column represents the scoring data obtained from a slice of your TMA. The name of the antibody, as specified in your original raw scoring workbook, used to stain that slice, appears at the top row of that column.
• In Row 2 is the EWEIGHT column, which defines the absolute weight each slice is given in the clustering. the default value for each slice is 1, indicating that each slice is given equal weight in the hierarchical clustering. You may also alter these values to some other number, prior to clustering.

If you had chosen to output in the K-M format, you will get the following instead:

K-M output file -- screenshot

Click on the image for a larger view of the picture.

• Column A provides information on the physical location of the spot within the TMA in the following format, s_a_c_r, where:
  s = sector number
  a = array number
  c = column number
  r = row number
  
• Column B provides information on the corresponding digital image filename for a particular spot. The nomenclature of this filename uses the numeric coding system used is based on the Bliss microscope system Bacus Laboratories Inc., consisting of seven numbers separated by underscores. This is covered in the Stainfinder walkthrough.
• Column C consists of the unique case identifier. It is labeled as "FP#" because the lookup files were generated from a FileMaker Pro database in the van de Rijn laboratory.
• Columns D-G (in this example; TMAs with a larger number of slices will occupy more columns in this range) contain the unmodified score data of the TMA.
• Column H (in this example; in other datasets, it will be the rightmost column in the file) contains the description for that spot and is formatted in the same way as the NAME column in the PCL output file.

You are now ready to proceed with hierarchical clustering or any other sort of statistical analysis you wish to perform on your TMA data.

Back to Step 1.

Step 3 - Clustering