The "Ok" button means that you will run the algorithm with the parameter values given in the "New value" column.

    The "Reset" button will set all the values in the "New value" column to those in the "Current value" column, invalidating all changes you made. The button is enabled only when you change at least one value.

    The "Cancel" button will return to the previous screen.

    A short explanatory text on the currently selected option is shown in the space immediately above (Macintosh) or below (PC, shown in the above figure) the list of parameters. Press Return or double-click the line to change the value of a parameter. If a parameter changes, the new value appear in the "New value column", and the field is marked by a red background (Macintosh) or a red arrow (PC).

The parameters have the following meaning.

a) Maximum number of markers per class

This number specifies the highest number of discriminant genes that will be taken into account for any class during the analysis. In case the multi-class analysis is selected (see below), this comprises markers for all class distinctions of a class.

b) Number of markers per couple

In a multi-class analysis, this specifies the maximum number of additional markers selected to distinguish among samples in a pair of classes .

c) Allow multi-class markers

When selected (Yes), ClassMarker is allowed to take into account marker genes that discriminate between a pair of classes against the others. This allows e.g. for detection of genes that behave significantly differently in a pair of diseases, such as two closely related cancers. If not selected (No), ClassMarker will only identify marker genes that discriminate a sample class against all the others.

d) Use k-NN classification

In the analysis, use the k-Nearest-Neighbours classification method instead of voting. Note that with k-NN, multi-class classification is impossible and the parameters b) and c) are ignored.