Package 'nipalsMCIA'

Title: Multiple Co-Inertia Analysis via the NIPALS Method
Description: Computes Multiple Co-Inertia Analysis (MCIA), a dimensionality reduction (jDR) algorithm, for a multi-block dataset using a modification to the Nonlinear Iterative Partial Least Squares method (NIPALS) proposed in (Hanafi et. al, 2010). Allows multiple options for row- and table-level preprocessing, and speeds up computation of variance explained. Vignettes detail application to bulk- and single cell- multi-omics studies.
Authors: Maximilian Mattessich [cre] , Joaquin Reyna [aut] , Edel Aron [aut] , Ferhat Ay [aut] , Steven Kleinstein [aut] , Anna Konstorum [aut]
Maintainer: Maximilian Mattessich <[email protected]>
License: GPL-3
Version: 1.5.2
Built: 2024-12-19 03:08:51 UTC
Source: https://github.com/bioc/nipalsMCIA

Help Index

Block-level preprocessing

Description

A function that normalizes an input dataset (data block) according to a variety of options. Intended to be used after column/row-level normalization.

Usage

block_preproc(df, block_preproc_method)

Arguments

df

dataset to preprocess (must be in data matrix form)
block_preproc_method

method which is used to normalize blocks, with options:

  • 'unit_var' FOR CENTERED MATRICES ONLY - divides each block by the square root of its variance

  • 'num_cols' divides each block by the number of variables in the block.

  • 'largest_sv' divides each block by its largest singular value.

  • 'none' performs no preprocessing

Value

the preprocessed dataset

Examples

df <- matrix(rbinom(15, 1, prob = 0.3), ncol = 3)
preprocessed_dataframe <- block_preproc(df,"unit_var")

block_weights_heatmap

Description

Function to plot heatmap of block score weights

Usage

block_weights_heatmap(mcia_results)

Arguments

mcia_results

MCIA results object returned from 'nipals_multiblock'

Details

Plotting function for heatmap of block score weights

Value

heatmap object containing the block weights as a heatmap

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                  plots = "none", tol = 1e-12)
block_weights_heatmap(mcia_results)

Centered Column Profile Pre-processing

Description

Converts data blocks into centered column profiles where each block has unit variance. Mimics the pre-processing in the Omicade4 package (Meng et al. 2014)

Usage

cc_preproc(df)

Arguments

df

the data frame to apply pre-processing to, in "sample" x "variable" format

Details

Performs the following steps on a given data frame:

  • Offsets data to make whole matrix non-negative

  • Divides each column by its sum

  • Subtracts (row sum/total sum) from each row

  • Multiplies each column by sqrt(column sum/total sum)

  • Divides the whole data frame by its total variance (the sqrt of the sum of singular values)

Value

the processed data frame

Examples

df <- matrix(rbinom(15, 1, prob = 0.3), ncol = 3)
preprocessed_dataframe <- cc_preproc(df)

Centered Column Profile Pre-processing

Description

Converts data blocks into centered column profiles where each block has unit variance. Mimics the pre-processing in the Omicade4 package (Meng et al. 2014)

Usage

col_preproc(df, col_preproc_method)

Arguments

df

the data frame to apply pre-processing to, in "sample" x "variable" format
col_preproc_method

denotes the type of column-centered preprocessing. Options are:

  • 'colprofile' Performs the following steps on a given data frame:

    • Offsets data to make whole matrix non-negative

    • Divides each column by its sum

    • Subtracts (row sum/total sum) from each row

    • Multiplies each column by sqrt(column sum/total sum)

  • 'standardized' centers each column and divides by its standard deviation.

  • 'centered_only' ONLY centers data

Details

Performs preprocessing on a sample/variable (row/column) level according to the parameter given.

Value

the processed data frame

Examples

df <- matrix(rbinom(15, 1, prob = 0.3), ncol = 3)
preprocessed_dataframe <- col_preproc(df, col_preproc_method = 'colprofile')

NCI-60 Multi-Omics Data

Description

A dataset of measurements of 12,895 mRNA, 537 miRNA, and 7,016 protein variables (columns) on 21 cancer cell lines (rows) from the NCI-60 cancer cell line database.

Value

Large list with 3 elements (one for each omic)

Source

Meng et. al, 2016 supplementary materials https://doi.org/10.1093/bib/bbv108

References

https://github.com/aedin/NCI60Example

Deflation via block loadings

Description

Removes data from a data frame in the direction of a given block loadings vector.

Usage

deflate_block_bl(df, bl)

Arguments

df

a data frame in "sample" x "variable" format
bl

a block loadings vector in variable space

Details

Subtracts the component of each row in the direction of a given block loadings vector to yield a ‘deflated’ data matrix.

Value

the deflated data frame

Examples

df <- matrix(rbinom(15, 1, prob = 0.3), ncol = 3)
block_loading <- rbinom(3, 1, prob = 0.3)
deflated_data <- deflate_block_bl(df, block_loading)

Deflation via global scores

Description

Removes data from a data frame in the direction of a given global scores vector.

Usage

deflate_block_gs(df, gs)

Arguments

df

a data frame in "sample" x "variable" format
gs

a global scores vector in sample space

Details

Subtracts the component of each column in the direction of a given global scores vector to yield a ‘deflated’ data matrix.

Value

the deflated data frame

Examples

df <- matrix(rbinom(15, 1, prob = 0.3), ncol = 3)
global_score <- rbinom(5, 1, prob = 0.3)
deflated_data <- deflate_block_gs(df, global_score)

Extract a list of harmonized data matrices from an MAE object

Description

Extract a list of harmonized data matrices for input into nipals_multiblock() from an MAE object

Usage

extract_from_mae(MAE_object, subset_data = "all", harmonize = TRUE)

Arguments

MAE_object

an MAE object containing experiment data for extraction colData field optional experiments should either be SummarizedExperiment, SingleCellExperiment, or RangedSummarizedExperiment classes
subset_data

  • 'all' use all experiments in MAE object

  • 'c(omic1,omic2,...)' list of omics from names(MAE_object)

harmonize

A boolean whether samples should be checked for duplicates

  • 'TRUE' (default) merges duplicate samples via the 'MultiAssayExperiment::mergeReplicates' function

  • 'FALSE' skips sample duplicate check - USE THIS FOR LARGE-SAMPLE DATASETS.

Value

List of harmonized data matrices for input into 'nipals_multiblock()'

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
NCI60_input = extract_from_mae(data_blocks_mae,subset='all')

Assigning colors to different omics

Description

Creates a list of omics and associated colors for plotting. The default palette was chosen to be color-blindness friendly.

Usage

get_colors(
  mcia_results,
  color_pal = scales::viridis_pal,
  color_pal_params = list()
)

Arguments

mcia_results

object returned from nipals_multiblock() function
color_pal

a function which returns color palettes (e.g. scales)
color_pal_params

list of parameters for the corresponding function

Value

List of omics with assigned colors

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                 plots = "none", tol = 1e-12)
colors_omics <- get_colors(mcia_results)

Assigning colors to different values of a metadata column

Description

Creates a list of metadata columns and associated colors for plotting. The default palette was chosen to be color-blindness friendly.

Usage

get_metadata_colors(
  mcia_results,
  color_col,
  color_pal = scales::viridis_pal,
  color_pal_params = list()
)

Arguments

mcia_results

object returned from nipals_multiblock() function
color_col

an integer or string specifying the column that will be used for color_col
color_pal

a function which returns color palettes (e.g. scales)
color_pal_params

list of parameters for the corresponding function

Value

List of metadata columns with assigned colors

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                               colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                  plots = "none", tol = 1e-12)
colors_omics <- get_metadata_colors(mcia_results, "cancerType",
                                    color_pal_params = list(option = "E"))

Computes the total variance of a multi-omics dataset

Description

Computes the total variances of all data blocks in a multi-omics dataset, intended for datasets that do not use 'CCpreproc'

Usage

get_tv(ds)

Arguments

ds

a list of multi-omics dataframes/matrices in "sample x variable" format

Value

the total variance of the dataset (i.e. sum of block variances)

Examples

data(NCI60)
tot_var <- get_tv(data_blocks)

global_scores_eigenvalues_plot

Description

Function to plot eigenvalues of scores up to num_PCs

Usage

global_scores_eigenvalues_plot(mcia_results)

Arguments

mcia_results

MCIA results object returned from 'nipals_multiblock'

Details

Plotting function for eigenvalues of scores up to num_PCs

Value

Displays the contribution plot using eigenvalues

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                             colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                 plots = "none", tol=1e-12)
global_scores_eigenvalues_plot(mcia_results)

Plotting a heatmap of global factors scores (sample v. factors)

Description

Plots a heatmap of MCIA global scores

Usage

global_scores_heatmap(
  mcia_results,
  color_col = NULL,
  color_pal = scales::viridis_pal,
  color_pal_params = list(option = "D")
)

Arguments

mcia_results

the mcia object matrix after running MCIA, must also contain metadata with columns corresponding to color_col
color_col

an integer or string specifying the column that will be used for color_col
color_pal

a list of colors or function which returns a list of colors
color_pal_params

a list of parameters for the color function

Value

ComplexHeatmap object

Perform biological annotation-based comparison

Description

Runs fgsea for the input gene vector

Usage

gsea_report(
  metagenes,
  path.database,
  factors = NULL,
  pval.thr = 0.05,
  nproc = 4
)

Arguments

metagenes

Vector of gene scores where the row names are HUGO symbols
path.database

path to a GMT annotation file
factors

vector of factors which should be analyzed
pval.thr

p-value threshold (default to 0.05)
nproc

number of processors to utilize

Value

data frame with the most significant p-value number of significant pathways

the selectivity scores across the given factors

NCI-60 Multi-Omics Metadata

Description

Metadata for the included multi-omics dataset, denoting the cancer type associated with each of the 21 cell lines.

Value

List with 21 elements

Source

Meng et. al, 2016 supplementary materials https://doi.org/10.1093/bib/bbv108

References

https://github.com/aedin/NCI60Example

NIPALS Iteration

Description

Applies one iteration stage/loop of the NIPALS algorithm.

Usage

nipals_iter(ds, tol = 1e-12, maxIter = 1000)

Arguments

ds

a list of data matrices, each in "sample" x "variable" format
tol

a number for the tolerance on the stopping criterion for NIPALS
maxIter

a number for the maximum number of times NIPALS should iterate

Details

Follows the NIPALS algorithm as described by Hanafi et. al. (2010). Starts with a random vector in sample space and repeatedly projects it onto the variable vectors and block scores to generate block and global loadings/scores/weights. The loop stops when either the stopping criterion is low enough, or the maximum number of iterations is reached. Intended as a utility function for 'nipals_multiblock' to be used between deflation steps.

Value

a list containing the global/block scores, loadings and weights for a given order

Examples

data(NCI60)
data_blocks <- lapply(data_blocks, as.matrix)
nipals_results <- nipals_iter(data_blocks, tol = 1e-7, maxIter = 1000)

Main NIPALS computation loop

Description

Applies the full adjusted NIPALS algorithm to generate block and global scores/loadings with the desired deflation method.

Usage

nipals_multiblock(
  data_blocks_mae,
  col_preproc_method = "colprofile",
  block_preproc_method = "unit_var",
  num_PCs = 10,
  tol = 1e-09,
  max_iter = 1000,
  color_col = NULL,
  deflationMethod = "block",
  plots = "all",
  harmonize = TRUE
)

Arguments

data_blocks_mae

a MultiAssayExperiment class object (with sample metadata as a dataframe in the colData attribute).
col_preproc_method

an option for the desired column-level data pre-processing, either:

  • 'colprofile' applies column-centering, row and column weighting by contribution to variance.

  • 'standardized' centers each column and divides by its standard deviation.

  • 'centered_only' ONLY centers data

block_preproc_method

an option for the desired block-level data pre-processing, either:

  • 'unit_var' FOR CENTERED MATRICES ONLY - divides each block by the square root of its variance

  • 'num_cols' divides each block by the number of variables in the block.

  • 'largest_sv' divides each block by its largest singular value.

  • 'none' performs no preprocessing

num_PCs

the maximum order of scores/loadings
tol

a number for the tolerance on the stopping criterion for NIPALS
max_iter

a number for the maximum number of times NIPALS should iterate
color_col

Optional argument with the column name of the 'metadata' data frame used to define plotting colors
deflationMethod

an option for the desired deflation method, either:

  • 'block' deflation via block loadings (for MCIA, default)

  • 'global' deflation via global scores (for CPCA)

plots

an option to display various plots of results:

  • 'all' displays plots of block scores, global scores, and eigenvalue scree plot

  • 'global' displays only global score projections and eigenvalue scree plot

  • 'none' does not display plots

harmonize

boolean whether or not samples should be checked for duplicates and re-ordered so that each row corresponds to the same sample across datasets. Set to FALSE to greatly reduce computation time on many samples (default = TRUE).

Details

Follows the NIPALS algorithm as described by Hanafi et. al. (2010). For each order of scores/loadings, the vectors are computed via the 'nipals_iter' function, then used to deflate the data matrix according to the desired deflation method. This process is repeated up to the desired maximum order of scores/loadings.

Value

a 'nipalsResult' object with the following fields:

  • 'global_scores' a matrix containing global scores as columns (NOT normalized to unit variance)

  • 'global_loadings' a matrix containing global loadings as columns

  • 'global_score_weights' a matrix of weights to express global scores as a combination of block scores. Has dimensions "num_Blocks" by "num_PCs"

  • 'eigvals' a matrix containing the eigenvalue for each computed global score.

  • 'block scores' a list of matrices, each contains the scores for one block

  • 'block loadings' a list of matrices, each contains the loadings for one block (w/ unit length)

  • 'block score weights' a matrix containing weights for each block score of each order used to construct the global scores.

  • 'col_preproc_method' the column preprocessing method used on the data.

  • 'block_variances' a list of variances of each block AFTER NORMALIZATION OPTION APPLIED

  • 'metadata' the metadata dataframe supplied with the 'metadata' argument. Note: overrides metadata present in any MAE class object.

Examples

data(NCI60)
   data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                                 colData=metadata_NCI60)
   NIPALS_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                       tol = 1e-12, max_iter = 1000,
                                       col_preproc_method = "colprofile",
                                       deflationMethod = "block")
   MCIA_results <- nipals_multiblock(data_blocks_mae, num_PCs = 4)
   CPCA_results <- nipals_multiblock(data_blocks_mae, num_PCs = 4,
   deflationMethod = 'global')

An S4 class to contain results computed with 'nipals_multiblock()'

Description

An S4 class to contain results computed with 'nipals_multiblock()'

Value

A NipalsResult object.

Slots

global_scores

A matrix containing global scores as columns.

global_loadings

A matrix containing global loadings as columns.

block_score_weights

A matrix containing block weights as columns.

block_scores

A list of matrices. Each matrix contains the scores as columns for a given block.

block_loadings

A list of matrices. Each matrix contains the loadings as columns for a given block.

eigvals

A list of singular values of the data matrix at each deflation step.

col_preproc_method

character for the column-level preprocessing method used. See 'col_preproc()'.

block_preproc_method

character for the block-level preprocessing method used. See 'block_preproc()'.

block_variances

A list of variances for each block.

metadata

A data frame of metadata originally passed into 'nipals_multiblock()'.

Accessor function for block loadings

Description

Retrieves the block loadings as a list of matrices from a 'NipalsResult' object, typically output from 'nipals_multiblock()'.

Usage

nmb_get_bl(nmb_object)

Arguments

nmb_object

A 'NipalsResult' object.

Value

a list of matrices containing block loadings.

Examples

data("NCI60")
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_out <- nipals_multiblock(data_blocks_mae, num_PCs = 10)
block_loadings<- nmb_get_bl(mcia_out)

Accessor function for block scores

Description

Retrieves the block scores as a list of matrices from a 'NipalsResult' object, typically output from 'nipals_multiblock()'.

Usage

nmb_get_bs(nmb_object)

Arguments

nmb_object

A 'NipalsResult' object.

Value

a list of matrices containing block scores.

Examples

data("NCI60")
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_out <- nipals_multiblock(data_blocks_mae, num_PCs = 10)
block_scores <- nmb_get_bs(mcia_out)

Accessor function for block score weights

Description

Retrieves the block score weights from a 'NipalsResult' object, typically output from 'nipals_multiblock()'.

Usage

nmb_get_bs_weights(nmb_object)

Arguments

nmb_object

A 'NipalsResult' object.

Value

a matrix containing the block score weights.

Examples

data("NCI60")
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                             colData=metadata_NCI60)
mcia_out <- nipals_multiblock(data_blocks_mae, num_PCs = 10)
block_score_weights <- nmb_get_bs_weights(mcia_out)

Accessor function for eigenvalues

Description

Retrieves the eigenvalues from a 'NipalsResult' object, typically output from 'nipals_multiblock()'.

Usage

nmb_get_eigs(nmb_object)

Arguments

nmb_object

A 'NipalsResult' object.

Value

a matrix containing the eigenvalues for all global scores.

Examples

data("NCI60")
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_out <- nipals_multiblock(data_blocks_mae, num_PCs = 10)
nipals_eigvals <- nmb_get_eigs(mcia_out)

Accessor function for global loadings

Description

Retrieves the global loadings as a matrix from a 'NipalsResult' object, typically output from 'nipals_multiblock()'.

Usage

nmb_get_gl(nmb_object)

Arguments

nmb_object

A 'NipalsResult' object.

Value

a matrix containing global loadings.

Examples

data("NCI60")
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                   colData=metadata_NCI60)
mcia_out <- nipals_multiblock(data_blocks_mae, num_PCs = 10)
global_loadings <- nmb_get_gl(mcia_out)

Accessor function for global scores

Description

Retrieves the global scores as a matrix from a 'NipalsResult' object, typically output from 'nipals_multiblock()'.

Usage

nmb_get_gs(nmb_object)

Arguments

nmb_object

A 'NipalsResult' object.

Value

a matrix containing global scores.

Examples

data("NCI60")
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                               colData=metadata_NCI60)
mcia_out <- nipals_multiblock(data_blocks_mae, num_PCs = 10)
global_scores <- nmb_get_gs(mcia_out)

Accessor function for metadata

Description

Retrieves the metadata from a 'NipalsResult' object, typically output from 'nipals_multiblock()'.

Usage

nmb_get_metadata(nmb_object)

Arguments

nmb_object

A 'NipalsResult' object.

Value

a dataframe containing metadata associated with the 'NipalsResult' object.

Examples

data("NCI60")
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_out <- nipals_multiblock(data_blocks_mae, num_PCs = 10)
nipals_metadata <- nmb_get_metadata(mcia_out)

Ranked global loadings dataframe

Description

Creates a dataframe with ranked loadings for a given factor

Usage

ord_loadings(
  mcia_results,
  omic = "all",
  factor = 1,
  absolute = FALSE,
  descending = TRUE
)

Arguments

mcia_results

object returned from nipals_multiblock() function
omic

choose an omic to rank, or choose 'all' for all, ((omic = "all", omic = "miRNA", etc.))
factor

choose a factor (numeric value from 1 to number of factors in mcia_results)
absolute

whether to rank by absolute value
descending

whether to rank in descending or ascending order

Value

ranked dataframe

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                  plots = "none", tol = 1e-12)
all_pos_1 <- ord_loadings(mcia_results = mcia_results, omic = "all",
                          absolute = FALSE, descending = TRUE, factor = 1)

Prediction of new global scores based on block loadings and weights

Description

Uses previously-computed block scores and weights to compute a global score for new data. Only validated for MCIA results, as CPCA loadings aren't compatible with un-deflated data.

Usage

predict_gs(mcia_results, test_data)

Arguments

mcia_results

an mcia object output by nipals_multiblock() containing block scores, weights, and pre-processing identifier.
test_data

an MAE object with the same block types and features as the training dataset. Feature and omic order must match 'bl'.

Details

Projects the new observations onto each block loadings vector, then weights the projection according to the corresponding block weights.

Value

a matrix of predicted global scores for the training data

Examples

data(NCI60)
   data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                                 colData=metadata_NCI60)
   mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 2)
   new_data <- data_blocks_mae # should update with a truly new dataset
   preds <- predict_gs(mcia_results, new_data)

projection_plot

Description

Function to generate a projection plot of MCIA results.

Usage

projection_plot(
  mcia_results,
  projection,
  orders = c(1, 2),
  block_name = NULL,
  color_col = NULL,
  color_pal = scales::viridis_pal,
  color_pal_params = list(option = "E"),
  legend_loc = "bottomleft",
  color_override = NULL,
  cex = 0.5
)

Arguments

mcia_results

MCIA results object returned from 'nipals_multiblock'
projection

of plot, with the following options

  • 'all' - scatter plot of two orders of global and block scores (aka factors).

  • 'global' - scatter plot of two orders of global scores only (aka factors).

  • 'block' - scatter plot of two orders of block scores only (aka factors) for given block.

orders

Option to select orders of factors to plot against each other (for projection plots)
block_name

Name of the block to be plotted (if 'projection = block' argument used).
color_col

an integer or string specifying the column that will be used for color_col
color_pal

a list of colors or function which returns a list of colors
color_pal_params

a list of parameters for the color function
legend_loc

Option for legend location, or "none" for no legend.
color_override

Option to override colors when necessary, helpful for projection = "global" or "block"
cex

Resizing parameter for drawing the points

Details

Plotting function for a projection plot.

Value

Displays the desired plots

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                  plots = "none", tol = 1e-12)
projection_plot(mcia_results, projection = "all", orders = c(1,2),
                color_col = "cancerType", legend_loc = "bottomright")

Create an MAE object from a list of data matrices and column data

Description

Create an MAE object from a set of data matrices and column data.

Usage

simple_mae(matrix_list, row_format = "feature", colData_input = NULL)

Arguments

matrix_list

named list of data matrices
row_format

for lists of data frames, indicates whether rows of datasets denote 'feature' (default) or 'sample'.
colData_input

a data frame containing sample metadata; sample names in the rownames should correspond to samples names in 'matrix_list'

Details

Requires that sample names match across experiments and are identical to primary names, will only convert data matrices to SummarizedExperiment class. If the data is more complex, please follow the guidelines for creating an MAE object outlined in 'help(MultiAssayExperiment)'

Value

List of harmonized data matrices for input into nipals_multiblock()

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks, row_format = "sample",
                              colData = metadata_NCI60)

Visualize ranked loadings

Description

Visualize a scree plot of loadings recovered from nipalsMCIA() output loadings matrix ranked using the ord_loadings() functions

Usage

vis_load_ord(
  mcia_results,
  omic,
  factor = 1,
  n_feat = 15,
  absolute = TRUE,
  descending = TRUE,
  color_pal = scales::viridis_pal,
  color_pal_params = list()
)

Arguments

mcia_results

object returned from nipals_multiblock() function
omic

name of the given omic dataset
factor

choose a factor (numeric value from 1 to number of factors in mcia_results)
n_feat

number of features to visualize
absolute

whether to rank by absolute value
descending

whether to rank in descending or ascending order
color_pal

a list of colors or function which returns a list of colors
color_pal_params

a list of parameters for the color function

Value

Plot in features for a factor by rank

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                  plots = "none", tol = 1e-12)
vis_load_ord(mcia_results, omic="mrna")

Visualize all loadings on two factor axes

Description

Visualize all loadings recovered from nipalsMCIA() output loadings matrix ranked using across two factor axes

Usage

vis_load_plot(
  mcia_results,
  axes = c(1, 2),
  color_pal = scales::viridis_pal,
  color_pal_params = list()
)

Arguments

mcia_results

object returned from nipals_multiblock() function
axes

list of two numbers associated with two factors to visualize
color_pal

a list of colors or function which returns a list of colors
color_pal_params

a list of parameters for the color function

Value

Plot of MCIA feature loadings for chosen axes

Examples

data(NCI60)
data_blocks_mae <- simple_mae(data_blocks,row_format="sample",
                              colData=metadata_NCI60)
mcia_results <- nipals_multiblock(data_blocks_mae, num_PCs = 10,
                                  plots = "none", tol = 1e-12)
vis_load_plot(mcia_results, axes = c(1, 4))