Package 'sketchR'

Title: An R interface for python subsampling/sketching algorithms
Description: Provides an R interface for various subsampling algorithms implemented in python packages. Currently, interfaces to the geosketch and scSampler python packages are implemented. In addition it also provides diagnostic plots to evaluate the subsampling.
Authors: Charlotte Soneson [aut, cre] , Michael Stadler [aut] , Friedrich Miescher Institute for Biomedical Research [cph]
Maintainer: Charlotte Soneson <[email protected]>
License: MIT + file LICENSE
Version: 1.1.0
Built: 2024-07-22 05:43:58 UTC
Source: https://github.com/bioc/sketchR

Help Index

Compare the compositions of a data set and a subset

Description

Plot the composition of a data set (e.g., the number of cells from each cell type) and contrast it with the corresponding composition of a subset.

Usage

compareCompositionPlot(
  df,
  idx,
  column,
  showPercentages = TRUE,
  fontSizePercentages = 4
)

Arguments

df

A data.frame-like object (such that df[[column]] works).
idx

A numeric vector representing the row indexes of df corresponding to the subset of interest. Can also be a named list of index vectors if multiple subsets are of interest.
column

A character scalar corresponding to a column of df and representing the variable for which the composition should be calculated.
showPercentages

Logical scalar, indicating whether relative frequencies of each category should be shown in the plot.
fontSizePercentages

Numerical scalar, indicating the font size of the relative frequencies, if showPercentages is TRUE.

Value

A ggplot object.

Author(s)

Charlotte Soneson

Examples

df <- data.frame(celltype = sample(LETTERS[1:5], 1000, replace = TRUE,
                                   prob = c(0.1, 0.2, 0.5, 0.05, 0.15)))
idx <- sample(seq_len(1000), 200)
compareCompositionPlot(df, idx, "celltype")

Run geosketch to subsample a matrix

Description

Perform geometric sketching with the geosketch python package.

Usage

geosketch(
  mat,
  N,
  replace = FALSE,
  k = "auto",
  alpha = 0.1,
  seed = NULL,
  max_iter = 200,
  one_indexed = TRUE,
  verbose = FALSE
)

Arguments

mat

m x n matrix. Samples (the dimension along which to subsample) should be in the rows, features in the columns.
N

Numeric scalar, the number of samples to retain.
replace

Logical scalar, whether to sample with replacement.
k

Numeric scalar or "auto", specifying the number of covering. If k = "auto" (the default), it is set to sqrt(nrow(mat)) for replace = TRUE and to N for replace = FALSE.
alpha

Numeric scalar defining the acceptable interval around k. Binary search halts when it obtains between k * (1 - alpha) and k * (1 + alpha) covering boxes.
seed

Numeric scalar or NULL (default). If not NULL, it will be converted to integer and passed to numpy to seed the random number generator.
max_iter

Numeric scalar giving the maximum iterations at which to terminate binary search in rare cases of non-monotonicity of covering boxes.
one_indexed

Logical scalar, whether to return one-indexed indices.
verbose

Locigal scalar, whether to print logging output while running.

Details

The first time this function is run, it will create a conda environment containing the geosketch package. This is done via the basilisk R/Bioconductor package - see the documentation for that package for troubleshooting.

Value

A numeric vector with indices to retain.

Author(s)

Charlotte Soneson, Michael Stadler

References

Hie et al (2019): Geometric sketching compactly summarizes the single-cell transcriptomic landscape. Cell Systems 8, 483–493.

Examples

x <- matrix(rnorm(500), nrow = 100)
geosketch(mat = x, N = 10, seed = 42)

Get names of geosketch functions

Description

Get names of geosketch functions

Usage

getGeosketchNames()

Value

A list of names of objects exposed in the geosketch module

Author(s)

Charlotte Soneson

Examples

getGeosketchNames()

Get names of scSampler functions

Description

Get names of scSampler functions

Usage

getScSamplerNames()

Value

A list of names of objects exposed in the scSampler module

Author(s)

Charlotte Soneson

Examples

getScSamplerNames()

Create diagnostic plot of Hausdorff distances

Description

Create diagnostic plot showing the Hausdorff distance between a sketch and the full data set, for varying sketch sizes. For reproducibility, seed the random number generator before calling this function using set.seed.

Usage

hausdorffDistPlot(
  mat,
  Nvec,
  Nrep = 5,
  q = 1e-04,
  methods = c("geosketch", "scsampler", "uniform"),
  extraArgs = list()
)

Arguments

mat

m x n matrix. Samples (the dimension along which to subsample) should be in the rows, features in the columns.
Nvec

Numeric vector of sketch sizes.
Nrep

Numeric scalar indicating the number of sketches to draw for each sketch size.
q

Numeric scalar in [0,1], indicating the fraction of largest minimum distances to discard when calculating the robust Hausdorff distance. Setting q=0 gives the classical Hausdorff distance. The default is 1e-4, as suggested by Hie et al (2019).
methods

Character vector, indicating which method(s) to include in the plot. Should be a subset of c("geosketch", "scsampler", "uniform"), where "uniform" randomly samples from input features with uniform probabilities.
extraArgs

Named list providing extra arguments to the respective methods (beyond the matrix and the sketch size). The names of the list should be the method names (currently, "geosketch" or "scsampler"), and each list element should be a named list of argument values. See the examples for an illustration of how to use this argument. Note that the seed argument, if provided to any of the methods, will be ignored (since it would imply providing the same seed for each repeated run of the sketching).

Value

A ggplot object.

Author(s)

Charlotte Soneson, Michael Stadler

References

Hie et al (2019): Geometric sketching compactly summarizes the single-cell transcriptomic landscape. Cell Systems 8, 483–493.

Song et al (2022): scSampler: fast diversity-preserving subsampling of large-scale single-cell transcriptomic data. bioRxiv doi:10.1101/2022.01.15.476407

Huttenlocher et al (1993): Comparing images using the Hausdorff distance. IEEE Transactions on Pattern Analysis and Machine Intelligence 15(9), 850-863.

Examples

## Generate example data matrix
mat <- matrix(rnorm(1000), nrow = 100)

## Generate diagnostic Hausdorff distance plot
## (including all available methods)
hausdorffDistPlot(mat, Nvec = c(10, 25, 50))

## Provide additional arguments for geosketch
hausdorffDistPlot(mat, Nvec = c(10, 25, 50), Nrep = 2,
                  extraArgs = list(geosketch = list(max_iter = 100)))

Run scSampler to subsample a matrix

Description

Perform subsampling with the scSampler python package.

Usage

scsampler(mat, N, random_split = 1, seed = 0)

Arguments

mat

m x n matrix. Samples (the dimension along which to subsample) should be in the rows, features in the columns.
N

Numeric scalar, the number of samples to retain.
random_split

Numeric scalar, the number of parts to randomly split the data into before subsampling within each part. A larger value will speed up computations, but give less optimal results.
seed

Numeric scalar, passed to scsampler to seed the random number generator.

Details

The first time this function is run, it will create a conda environment containing the scSampler package. This is done via the basilisk R/Bioconductor package - see the documentation for that package for troubleshooting.

Value

A numeric vector with indices to retain.

Author(s)

Charlotte Soneson, Michael Stadler

References

Song et al (2022): scSampler: fast diversity-preserving subsampling of large-scale single-cell transcriptomic data. bioRxiv doi:10.1101/2022.01.15.476407

Examples

x <- matrix(rnorm(500), nrow = 100)
scsampler(mat = x, N = 10)