doSVD: Perform Truncated Singular Value Decomposition (SVD)

In khazum/ccImpute: ccImpute: an accurate and scalable consensus clustering based approach to impute dropout events in the single-cell RNA-seq data (https://doi.org/10.1186/s12859-022-04814-8)

Perform Truncated Singular Value Decomposition (SVD)

Description

Computes a truncated SVD on a matrix using the implicitly restarted Lanczos bidiagonalization algorithm (IRLBA).

Usage

doSVD(x, svdMaxRatio = 0.08, nCeil = 2000, nCores)

Arguments

x	A numeric matrix to perform SVD on.
svdMaxRatio	(Optional) The maximum proportion of singular vectors used for generating subsets (default: 0.08).
nCeil	(Optional) The maximum number of cells used to compute the proportion of singular vectors (default: 2000).
nCores	The number of cores to use for parallel processing.

Details

This function utilizes the 'irlba' function from the 'irlba' package to efficiently calculate the truncated SVD of the input matrix 'x'. The returned matrix contains 'nv' right singular vectors, which are often used for dimensionality reduction and feature extraction in various applications.

Value

A matrix containing the right singular vectors of 'x'.

Examples

library(scater)
library(splatter)

sce <- splatSimulate(group.prob = rep(1, 5)/5, sparsify = FALSE, 
        batchCells=100, nGenes=1000, method = "groups", verbose = FALSE, 
        dropout.type = "experiment")
sce <- logNormCounts(sce)
cores <- 2
logX <- as.matrix(logcounts(sce))
w <- rowVars_fast(logX, cores)
corMat <- getCorM("spearman", logcounts(sce), w, cores)
v <- doSVD(corMat, nCores=cores)

khazum/ccImpute documentation built on July 26, 2024, 1:13 a.m.