
Entering Data
One can use the Data tab to enter the data. The appropriate data is in comma
separated values (csv) format with separated columns for the response, ID, dose,
etc. Once the data is uploaded, different panels will appear on the screen to
specify different aspects of the data. In the following, we briefly explain each
of the panels in the Data tab.
- Select the name: Here the column with the names of subjects should be
selected, it could be for example, a column with the compound names.
- Select the ID: Here the column withe compound ID should be selected. This
ID will be used in the course of clustering, modeling, etc.
- Select the dose: Here the column indicating the dose should be selected.
Note that, for the computation restrictions, this dose should be a numeric
value. So even if you have low, medium, high doses, please change them with
the actual numeric dose before uploading the data file.
- Select the response: The column with the response to be modeled should be
specified here.
- Select the added covariates: It is possible that in some cases you want
to include other covariates other than the dose in your model. You may select
all of them here (one by one), note that they will be entered in the model in
a linear way.
- Select the number of digits to be displayed: When displaying the data, the
default is 2 digits, but here you can increase or decrease them.
- Is the dose-response pattern known to be monotone? If we already know our
dose-response pattern SHOULD be monotone, it can be used for the analysis, so
here this should be identified.
Estimating ED50
Once the data is uploaded and every needed aspect is specified, one can
proceed to estimating the effective dose. This can be done via the Modelling
tab. There one may see these options:
- Select the effective dose to estimate (EDp): The default is 0.5, i.e.,
ED50, but one can specify any other p here also.
- Which transform should be applied to the response? In case the response
needs to be transformed, one can specify it here, the available transforms
are:
- no transformation (none)
- natural logarithm (log)
- square root (sRoot)
- qubic root (qRoot)
- Box-Cox (boxcox)
- Enter the toxic dose threshold: In some cases, the scientist has an idea
about a toxic dose threshold, this dose can be entered here, this way an
extra column will be created with the final results indicating whether each
compound is toxic or not (based on estimated EDp).
- Estimate EDp: Once everything is specified, one may press the “Estimate
EDp” button to begin the model fitting procedure. Depends on the amount of
data it may take several minutes.