AddLayout: Add 2-dimensional layout to a Benchmark object

In davnovak/SingleBench: Framework for benchmarking and optimising dimension-reduction and clustering pipelines in expression data analysis

Description

Applies a projection tool to produce a 2-dimensional layout of all input data, for visualisation purposes downstream.

Usage

AddLayout(
  benchmark,
  idx.subpipeline,
  idx.n_param,
  method,
  idcs_training,
  coordinates,
  hpc,
  qsub_config,
  ask_overwrite,
  seed,
  verbose
)

Arguments

benchmark	object of type Benchmark object created using the Benchmark constructor and successfully evaluated using Evaluate
idx.subpipeline	optional integer: sub-pipeline index to identify existing latent-space projection to use. Default value is NULL
idx.n_param	optional integer: n-parameter iteration index to identify existing latent-space projection to use. Default value is NULL
method	optional object of class WrapperWithParameters, generated by the function Fix, that is set to produce a 2-dimensional layout of input data associated with the benchmark object
idcs_training	optional vector of integers: if input dataset consists of multiple expression matrices (corresponding to multiple samples), which of them should be used for training the dimension-reduction model? Default value is NULL, which causes all of them to be used
coordinates	optional numeric matrix: row-wise coordinates of a 2-dimensional layout to use. If specified, this parameter overrides idx.subpipeline, idx.n_param, wrapper and params
ask_overwrite	logical: if a 2-dimensional layout is already available, should the user be asked whether to overwrite it? Default value is TRUE
seed	optional numeric value: value random seed to be used prior to the wrapper call (or NULL to avoid setting a seed). Default value is 1
verbose	logical: should progress messages be printed during evaluation? Default value is TRUE

Details

During the evaluation, a 2-dimensional latent-space projection of the input dataset might have been created already. In that case, you can specify, by sub-pipeline and n-parameter iteration index, the projection to use as layout (instead of generating a new one). Moreover, you can specify a more-than-2-dimensional projection, in which case the first two column of its coordinate matrix are used (this might be sensible when using PCA or diffusion maps).

Finally, if you want to use an existing coordinate matrix instead, you can pass it as a parameter to this function and it will be used as layout.

See Also

• Plot: lets you produce a plot of your choosing, visualising results of a previously evaluated benchmark pipeline.

davnovak/SingleBench documentation built on Dec. 19, 2021, 9:10 p.m.