Contents
Experiment Results Files
The database generates .pcl files when data are retrieved. After clustering, a .cdt file is generated, and in addition, .gtr and .atr files may also be generated. The complete dataset without any processing can also be downloaded as an Excel file.
pcl file format
The pcl file is a tab-delimited pre-clustering file. The first three columns are as follows:
- Name
- This column will contain a name ascribed to the entity on that
row, such as ORF name, or CLONEID. The column itself can be named
anything, but by convention is named YORF when it contains yeast ORF
names, CLID when it contains cloneids, and LUID when it contains
LUIDs. These name are simply so that after clustering, the Treeview software can
use the contents of the column in URLs without additional
configuration. This column MUST contain some text on every row.
- Description
- This column can contain descriptive information about the
entity, eg process or function, or gene symbol. It too can be named
anything. It can optionally be left blank, but the column itself must
be present.
- GWEIGHT
- This column allows you to weight genes with
different weights, for instance if a gene appears on an array twice,
you may want to give them a weight of 0.5 each. For the most part
people leave this column with a value of 1 for every gene. This
column must be present, and each row must have an entry.
In addition the file must begin with the following two rows:
- Row 1
- This contains the column headers as described above for
columns 1, 2 and 3, then contains the experiment names for all the
data columns that exist in the file. Each data column must have a
text entry as a name for that column.
- Row 2
- This is the EWEIGHT row. The entry in the first column for
this row should say EWEIGHT, then for each experiment, there should be
an EWEIGHT value. This will usually be 1, but if the same experiment
is duplicated twice, you may want to give these repeats an EWEIGHT of
0.5.
The remaining cells in the file contain the actual data, such that the row and column specifies to which gene and which experiment a particular piece of data corresponds. If you had modified, or created your pcl file in Excel, it would look something like this:
You should then choose Save As... from the File menu, and elect to the the file as type Text (Tab delimited), as indicated below:
- The Data
In general the pcl file will contain log-transformed data, which is needed for clustering to work properly.
- Name
cdt file format
When you cluster a .pcl file you will generate a .cdt (clustered data table) file, which will contain the original data, but reordered, to reflect the clustering. In addition, if you clustered by genes, you will get a .gtr file (gene tree), and if you clustered by experiments you will get a .atr file (array tree). These tree file reflect the history of how the cluster was built, and can be used to contruct how the tree(s) should look.
gtr file format
The .gtr (gene tree) file records the order in which the genes (rows) were joined during clustering.
atr file format
The .atr (array tree) file records the order in which the arrays (columns) were joined during clustering.
