scde (version 2.0.1)

pagoda.top.aspects: Score statistical significance of gene set and cluster overdispersion

Description

Evaluates statistical significance of the gene set and cluster lambda1 values, returning either a text table of Z scores, etc, a structure containing normalized values of significant aspects, or a set of genes underlying the significant aspects.

Usage

pagoda.top.aspects(pwpca, clpca = NULL, n.cells = NULL,
  z.score = qnorm(0.05/2, lower.tail = FALSE), return.table = FALSE,
  return.genes = FALSE, plot = FALSE, adjust.scores = TRUE,
  score.alpha = 0.05, use.oe.scale = FALSE, effective.cells.start = NULL)

Arguments

pwpca

output of pagoda.pathway.wPCA()

clpca

output of pagoda.gene.clusters() (optional)

n.cells

effective number of cells (if not provided, will be determined using pagoda.effective.cells())

z.score

Z score to be used as a cutoff for statistically significant patterns (defaults to 0.05 P-value

return.table

whether a text table showing

return.genes

whether a set of genes driving significant aspects should be returned

plot

whether to plot the cv/n vs. dataset size scatter showing significance models

adjust.scores

whether the normalization of the aspect patterns should be based on the adjusted Z scores - qnorm(0.05/2, lower.tail = FALSE)

score.alpha

significance level of the confidence interval for determining upper/lower bounds

use.oe.scale

whether the variance of the returned aspect patterns should be normalized using observed/expected value instead of the default chi-squared derived variance corresponding to overdispersion Z score

effective.cells.start

starting value for the pagoda.effective.cells() call

Value

• if return.table = FALSE and return.genes = FALSE (default) returns a list structure containing the following items:
  • xv
  {a matrix of normalized aspect patterns (rows- significant aspects, columns- cells}
• xvw
• { corresponding weight matrix }
• gw
• { set of genes driving the significant aspects }
• df
• { text table with the significance testing results }

Examples

data(pollen)
cd <- clean.counts(pollen)
knn <- knn.error.models(cd, k=ncol(cd)/4, n.cores=10, min.count.threshold=2, min.nonfailed=5, max.model.plots=10)
varinfo <- pagoda.varnorm(knn, counts = cd, trim = 3/ncol(cd), max.adj.var = 5, n.cores = 1, plot = FALSE)
pwpca <- pagoda.pathway.wPCA(varinfo, go.env, n.components=1, n.cores=10, n.internal.shuffles=50)
tam <- pagoda.top.aspects(pwpca, return.table = TRUE, plot=FALSE, z.score=1.96)  # top aspects based on GO only