weight.ref.pca: Weight sample difference from a set of reference samples

In jmonlong/PopSV: Population-based detection of structural variants from Read-Depth signal

Description

Compute weight that represent how similar samples are to the reference samples, from the PCA. These weights could be used to guide a better normalization or adjustment of the stringency for the calls. Indeed, some subtle 'abnormal' signal detected might be due to fact that a sample is globally different from the reference samples, hence a stronger normalization and/or a more stringent calling is advised. STILL IN DEVELOPMENT.

Usage

weight.ref.pca(pca.mat, ref.samples, plot = FALSE, output.dist = FALSE)

Arguments

pca.mat	a matrix with the principal components as columns and the sample names as row names. The first two components are used.
ref.samples	the names of the samples used as reference.
plot	should some graphs be displayed. Default is FALSE.
output.dist	should the distance to the reference sample centroid be outputed instead of the weights. Default is FALSE.

Value

a vector with the weights for each sample.

Author(s)

Jean Monlong

jmonlong/PopSV documentation built on Sept. 15, 2019, 9:29 p.m.