scNPF.fus: Similarity matrix Construction

In BMILAB/scNPF: Pre-processing of single-cell RNA-seq data

Description

constructs a sample-similarity network for each propagated expression matrix and then integrates these networks into a single cell-cell similarity network based on a nonlinear combination method.

Usage

scNPF.fus(data, K = 20, alpha = 0.5, T = 10)

Arguments

data	A list stored network-smoothed gene expression matrix with two length.
K	Number of neighbors in K-nearest neighbors part of the algorithm.
T	Number of iterations for the diffusion process.
alpah	Variance for local model.

Value

A consensus similarity matrix.

Examples

##Loading example data
load(system.file("data","yan.Rdata",package = "scNPF"))
exp.data <- yan$data[1:2000,]

##Smoothed gene expression based on priori mode
##Using String network.
load(system.file("data","string.Rdata",package = "scNPF"))
string.data <- scNPF.pro(x=exp.data, network=string)

##Smmothed gene expression based on context mode
context.data<- scNPF.pro(x=exp.data, network="context",qt.gene=0.9,qt.cell=0.9)

##Construction a cell-by-cell similarity matrix.
similarity <- scNPF.fus(data=list(string=string.data,context=context.data))

BMILAB/scNPF documentation built on Nov. 19, 2020, 1:41 a.m.