procsom: Processing of self-organizing maps for pseudo-temporal...

In dgrun/FateID: Quantification of Fate Bias in Multipotent Progenitors

Processing of self-organizing maps for pseudo-temporal expression profiles

Description

This function processes the self-organizing maps produced by the function getsom.

Usage

procsom(s1d, corthr = 0.85, minsom = 3)

Arguments

s1d	output of function getsom
corthr	correlation threshold, i. e. a real number between 0 and 1. The z-score of the average normalized pseudo-temporal expression profiles within each node of the self-organizing map is computed, and the correlation of these z-scores between neighbouring nodes is computed. If the correlation is greater than corthr, neighbouring nodes are merged. Default value is 0.85.
minsom	positive integer number. Nodes of the self-organizing map with less than minsom transcripts are discarded. Default value is 3.

Value

A list of the following seven components:

k	vector of Pearson's correlation coefficient between node i and node i+1 of the populated nodes of the self-organizing map.
nodes	vector with assignment of genes to nodes of the final self-organizing map (after merging). Components are node numbers and component names are gene IDs.
nodes.e	data frame with average normalized pseudo-temporal expression profile for each node, ordered by node number.
nodes.z	data frame with z-score transformed average normalized pseudo-temporal expression profile for each node, ordered by node number.
all.e	data frame with normalized pseudo-temporal expression profile for all genes in the self-organizing map, ordered by node number.
all.z	data frame with z-score transformed normalized pseudo-temporal expression profile for all genes in the self-organizing map, ordered by node number.
all.b	data frame with binarized pseudo-temporal expression profile for all genes in the self-organizing map, ordered by node number. Expression is 1 in cells with z-score > 1 and -1 in cells with z-score < -1, and 0 otherwise.

Examples

x <- intestine$x
y <- intestine$y
v <- intestine$v

tar <- c(6,9,13)
fb <- fateBias(x,y,tar,z=NULL,minnr=5,minnrh=10,nbfactor=5,use.dist=FALSE,seed=NULL,nbtree=NULL)
dr <- compdr(x,z=NULL,m="cmd",k=2,tsne.perplexity=30)
pr <- prcurve(y,fb,dr,k=2,m="cmd",trthr=0.4,start=NULL)
n <- pr$trc[["t6"]]
fs  <- filterset(v,n,minexpr=2,minnumber=1)
s1d <- getsom(fs,nb=1000,alpha=.5)
ps <- procsom(s1d,corthr=.85,minsom=3)

dgrun/FateID documentation built on June 20, 2022, 12:57 p.m.