runDiffusionMaps: Dimentionality Reduction by Diffusion Maps Algorithm

In r3fang/SnapATAC: Single Nucleus Analysis Package for ATAC-Seq

Dimentionality Reduction by Diffusion Maps Algorithm

Description

This function takes a snap obj as input and runs diffusion maps for dimentionality reduction.

Usage

runDiffusionMaps(obj, input.mat = c("bmat", "pmat"), num.eigs = 20)

Arguments

obj	A snap obj
input.mat	Input matrix c("bmat", "pmat").
num.eigs	Number of eigenvectors to be computed [20].

Details

Diffusion Maps algorithm, a nonlinear dimensionality reduction technique that discovers low dimensional manifolds within high-dimensional datasets by performing harmonic analysis of a random walk constructed over the data to identify nonlinear collective variables containing the predominance of the variance in the data. We choose diffusion maps because it is highly robust to noise and perturbation, making it particuarly suited for analyzing sparse scATAC-seq dataset.

Examples

data(demo.sp);
demo.sp = makeBinary(demo.sp);
demo.sp = runDiffusionMaps(
obj=demo.sp, 
input.mat="bmat", 
num.eigs=20
);

r3fang/SnapATAC documentation built on March 29, 2022, 4:33 p.m.