compute.pts: compute.pts

In staaln/pathtracer:

compute.pts

Description

Computes PathTracer deregulations scores based on gene expressions for a set of genes belonging to a particular pathway, and a set of samples, e.g normal and tumor samples.

Usage

compute.pts(dat, reference, ncomp = 4, pathwayindex)

Arguments

dat	Matrix of gene expression data with rows corresponding to genes and columns to samples.
reference	Bolean vector indicating which of the columns in dat that are reference samples (e.g normal samples).
ncomp	Number of principal components.
pathwayindex	Integer indicating which pathway the set of input genes belongs to.

Details

The main PathTracer function. The input is a matrix of gene expressions for a particular pathway and a set of samples, with rows representing genes, and colums samples. The function first calculates the principal curve for the first ncomp principal components and projects all samples onto the principal curve. The PathTracer deregulation score (PTS) for each sample is calculated as the Euclidean distance from the sample's projection onto the principal curve to the reference point (defined as the projection onto the curve for the reference sample with median distance along the curve to the start of the curve). See Nygard et al (2019), for furhter details.

Value

pts Vector of PathTracer deregulation scores

pds Vector of deregulation scores based on distance along the curve (similar to the pathifier algorithm).

auc.pts Area under the ROC based on pts values, where reference samples are treated as negatives and non-reference samples as positives.

auc.pds Area under the ROC based on pds values, where reference samples are treated as negatives and non-reference samples as positives.

data Normalized expression matrix.

ref Same as input variable reference.

res Output from principal_curve.

xcen Reference point.

v Matrix whose colums contain the left singular values of the normalized data matrix.

d Vector containing the singular values of the normalized data matrix.

pathwayindex Same as input pathwayindex

References

Nygard S, Lingjaerde OC, Caldas C, Hovig E, Borresen-Dahle, Helland Aa, Haakensen V. "PathTracer: High-sensitivity detection of differential pathway activity in tumours". Submitted.

staaln/pathtracer documentation built on March 17, 2022, 7:44 p.m.