scLINE: Integrated networks embedding for scRNA-seq data

In BMILAB/scLINE: A tool for R to perform low-dimensional representation of scRNA-seq using the scLINE model

Description

A tool of dimension reduction for scRNA-seq data combined with multiple gene networks embedding model

Usage

scLINE(mat, network, L, K = 5, T, rho = 0.025, Random_weight)

Arguments

mat	A matrix of scRNA-seq data in which rows represent gene columns representing cells
network	A list of gene networks with 3 columns: geneID, geneID, score
L	Dim of the low-dimensional representations
K	Number of negative samples, the default is 5
T	Number of iterations
rho	The initial learning rate, the default is 0.025
Random_weight	True of False,Whether to add random weights to the network model

Value

the low-dimensional matrix of input data

Examples

load(system.file("data","Usoskin.Rdata",package = "scLINE"))
load(system.file("data","ppi.Rdata",package = "scLINE"))
load(system.file("data","humannet.Rdata",package = "scLINE"))
exp_mat<-Usoskin$rawdata
gene_network<-list(ppi = ppi,humannet = humannet)
lowdim_list<-scLINE(exp_mat, gene_network, L = 20, K = 5, T = 1e7, rho = 0.025, Random_weight = TRUE)

BMILAB/scLINE documentation built on Oct. 30, 2019, 5:15 a.m.