CLI reference

domainator build_projection

Project a matrix into 2D or 3D with UMAP, t-SNE, or PCA.

Generates 2d or 3d umap, tsne, or pca projections from matrices

output can be an interactive HTML plot, coordinates, or Cytoscape xgmml files.

Give at least one of these outputs: --xgmml_out, --html_out, --dvb_out, --coords_out. - as an input or output means standard input or output, where the option says it is allowed.

Option	Value	Default	Description
-v, --verbose			show debug output, and a traceback for errors.
-q, --quiet			show errors only: no progress messages or warnings.
-i, --input	FILE	required	The matrix that you want to project into 2D or 3D space, rows are observations (typically gene names), columns are features (typically similarity to other genes, or hmm profile scores). Format can be tab-separated text or Domainator hdf5.
--algorithm	{pca,tsne,umap,coords}	pca	Which projection algorithm to use.
--metadata	FILE ...		tab separated files of sequence metadata.
--color_by	STRING		Color the points in the HTML plot and the xgmml by this column of the metadata table.
--color_table	FILE		tab separated file with two columns and no header, columns are: annotation, hex color. For example: CCDB cc0000
--color_table_out	FILE		tab separated file with two columns and no header, columns are: annotation, hex color. Written after the color table is updated with new colors, for example if using --color_by, but not supplying an external color table.
--palette_offset	INTEGER	0	Skip this many colors at the start of the palette before assigning colors to --color_by values. For example, if a plot colored by one column used the first 5 colors, --palette_offset 5 colors another column without reusing them. Ignored with --color_table.
--label_by	STRING		Label the points in the html viewer by this metadata column at first.
--categorical	STRING ...		Numeric metadata columns (for example integer cluster numbers) that the html viewer should color as discrete categories instead of a gradient.
--name	STRING		The projection's name, shown in the html viewer. Defaults to the html (or dvb) output's file name.
-o, --xgmml_out	FILE		Write the projection to this Cytoscape xgmml file.
--html_out	FILE		Write the Domainator Viewer, showing the projection, to this HTML file. It selects points into a metadata table and saves pictures as SVG or PNG.
--dvb_out	FILE		Write the gzip-compressed JSON Domainator Viewer bundle (.dvb) to this file. The viewer opens it, and saves sessions in the same format.
--coords_out	FILE		Write the projected coordinates to this tab-separated file.
--pca_args	STRING		keyword arguments for the pca implementation, as key=value pairs in one quoted string, for example --pca_args="n_components=2". Values are read as Python literals where possible.
--tsne_args	STRING		keyword arguments for the tsne implementation, as key=value pairs in one quoted string, for example --tsne_args="n_components=2". Values are read as Python literals where possible.
--umap_args	STRING		keyword arguments for the umap implementation, as key=value pairs in one quoted string, for example --umap_args="n_components=2". Values are read as Python literals where possible.
--scale	NUMBER	1000	scale the largest axis so that the difference between the largest and smallest value is this.
--three_d			if set, then makes 3d projections. default: 2d projections.

Every command also takes --config FILE (options from a JSON or YAML file, keyed by option name) and --print_config.