Up: Component summary Component

HeatMapReport

Clusters samples hierarchically and draws the corresponding dendrograms and heat map. The clustering is determined by the input matrix and the clustering parameter.

The size and character expansion factor of row and column names, color scheme and number of colors of the heat map, and the image margin sizes can all be adjusted.

The component outputs a LaTeX report fragment.

Version 3.0
Bundle tools
Categories
Authors Riku Louhimo (Riku.Louhimo@Helsinki.FI), Amjad Alkodsi (Amjad.Alkodsi@Helsinki.FI)
Issue tracker View/Report issues
Requires R ; pheatmap (R-package)
Source files component.xml HeatMap.r
Usage Example with default values

Inputs

Name Type Mandatory Description
in Matrix Mandatory Numeric data matrix.
rowLabels CSV Optional Annotation for matr rows. Unique colors are assigned to each unique group in label matrix and plotted next to the dendrogram. Values in rowLabels do not have to be in the same order as rows in matr input. Annotation column names cannot be names 'Col1', 'Col2',..., 'ColN' due to a bug in the pheatmap library.
colLabels CSV Optional Annotation for matr columns. Unique colors are assigned to each unique group in label matrix and plotted beneath the dendrogram. Values in colLabels do not have to be in the same order as columns in matr input. Annotation column names cannot be names 'Col1', 'Col2',..., 'ColN' due to a bug in the pheatmap library.

Outputs

Name Type Description
out Latex Clustering report.

Parameters

Name Type Default Description
caption string "" Image caption.
clusterMethod string "average" Hierarchical clustering method to be used. One of 'average', 'complete', 'ward', 'single', 'mcquitty', 'median' or 'centroid'. See the hclust R function help for a complete description of the alternatives. The same method applies to both columns and rows.
colorScheme string "colorRampPalette(c('blue','white','red'))((256))" Color palette used for the heatmap hues. One of 'cm', 'heat', 'topo', 'terrain', 'redblue' and 'redgreen'; or an R expression producing a vector of colors. For the R expression, the number of colors is set as specified by each function.
custom string "" A comma separated list of 'parameter=value' pairs that are passed to pheatmap AS THEY ARE. This allows customising the heatmap command. Parameters and values are not checked so use with caution.
distanceMetric string "euclidean,euclidean" Distance measure to be used in computing distances from the input matrix for the clustering. One of 'pcc', 'euclidean', 'manhattan', 'maximum', 'canberra', 'binary', 'minkowski', or 'none' if no reordering should be done. The parameter is a comma separated list of lenght two. The first value defines the metric for rows and the second value for columns. Setting either value 'NA' disables clustering and dendrogram for rows or columns. The special value 'NA,NA' disables clustering for both rows and columns. The default metric is the Euclidian distance.
drawColnames boolean true Enable or disable column names in the plot.
drawLegends boolean false Enable or disable legends for row and column labels in the plot.
drawRownames boolean true Enable or disable row names in the plot.
fontSizeCol float 10.0 Font size for column names.
fontSizeRow float 10.0 Font size for row names.
marginCol int 7 The margin size on the x-axis (input matrix columns).
marginRow int 4 The margin size on the y-axis (input matrix rows).
plotHeight int 5 The height of graphic region in inches.
plotWidth int 10 The width of graphic region in inches.
scale string "none" Indicates how centering and scaling will be performed. The possible values are "none","row", or "column". By default, the distance matrix is calculated from the scaled values (see 'visualScales' parameter).
sectionTitle string "Sample clusters" Title of the LaTeX section.
sectionType string "subsection" Type of LaTeX section: usually one of section, subsection or subsubsection. If empty, no section statement is written.
visualScales boolean false Indicates whether the distance matrix is calculated from scaled or unscaled values. This is useful especially when the data need to be clustered based on the raw values but scaling creates more a visually cleared figure.
width float 16 Absolute width of the diagram in centimeters.

Test cases

Test case Parameters IN
in
IN
rowLabels
IN
colLabels
OUT
out
case1 (missing) in (missing) (missing) out
case2 properties in (missing) (missing) out

clusterMethod = complete,
distanceMetric = NA,NA,
colorScheme = rev(heat.colors(256)),
marginRow = 5,
marginCol = 10,
fontSizeCol = 9,
fontSizeRow = 9,
width = 15,
drawRownames = false,
drawColnames = false

case3 properties in rowLabels colLabels (missing)

distanceMetric = euclidean,euclidean,
drawLegends = true,
drawRownames = true,
drawColnames = true,
marginRow = 4

case4 properties in rowLabels colLabels (missing)

distanceMetric = euclidean,euclidean,
drawLegends = true,
drawRownames = true,
drawColnames = true,
marginRow = 4,
custom = cutree_rows=2,cutree_cols=2

case5 properties in rowLabels colLabels (missing)

distanceMetric = euclidean,euclidean,
drawLegends = true,
drawRownames = true,
drawColnames = true,
marginRow = 4,
visualScales=true


Generated 2018-12-11 07:42:07 by Anduril 2.0.0