Map Reflection Tool

Reflection is defined as an operation to transition from one map to another via a selection of nodes. E.g. given a group of samples on a Sample Map, mRNA reflection allows the user to view genes on a Gene Map whose mRNA are up or down regulated in those samples, and visa versa. Technical details are described at the bottom of this page.

Reflecting a set of data from one map onto a companion map may be performed by clicking on the Tools menu then clicking on Map Reflection.

The resulting window, Reflect on Another Map is described below.

Not all maps have reflection maps. When you click on this menu option and the current map does not have the required data, a message will be displayed indicating the lack of data.

Input Descriptions

Selection : a binary attribute, possibly one you have created by selecting a group of nodes using an option from the Select menu. More information on selecting nodes is at Select Nodes . More information on binary attribute data types is at Attribute Data Types .

Map ID : the map ID of the target map upon which the data will be reflected. This list only contains companion maps that have information to display related to the current map.

Layout : the layout of the map ID upon which the data will be reflected. This list only contains layouts of the target map ID that have information to display related to the current map.

Click on the Reflect button to start up the reflection on the target map. After computations are complete a new browser tab will be opened with the target map displaying the reflected data.

Technical Details

Currently, reflection is an operation that transitions from columns<->rows of a data matrix. The data matrix is most often the same data matrix that is used to generate the matching data type layout. Before this operation is preformed, the data matrix undergoes a CLR transform (in all cases except for SCNV). The CLR transform is the euclidean distance of the column-wise and row-wise z-scores of each cell in the data matrix. This operation is intended to help with symmetry, i.e. if you reflect from a group of samples to a group of genes then reflect back to samples using the “high” reflection category from genes, you should generally be returned the original sample selection.

For mRNA, miRNA, Methylation, RPPA, a t-test is preformed and the top and bottom ranking t-statistics are highlighted.

For SCNV, gistic-like thresholds are averaged and the top and bottom ranking averages are highlighted.

Note: If a user has their own instance of TumorMap and would like to implement a reflection operation other than the ones listed above, please contact the developers. In general any python implementation whose input is column/row names and whose outputs is a ranking of row/column names could be used as a reflection operation.