Up: Component summary Component

ExprMixtureModel

Provides a mixture model fit of two normal distributions for the given genes. Each gene is processed individually.

Output columns:

gene
identifier of the expression profile
normality
P-value of associated with the Shapiro-Wilk normality test
loglik
log likelihood of the mixture model
m1mean
mean of the first Gaussian component
m2mean
mean of the second Gaussian component
m1sd
standard deviation of the first Gaussian component
m2sd
standard deviation of the second Gaussian component
m1prop
proportion of samples captured by the first Gaussian component
m2prop
proportion of samples captured by the second Gaussian component
selected
expression profile satisfies the given thresholds and is plotted
mi
normalized mutual information score

Version 1.4
Bundle microarray
Categories DEG Quality Control
Authors Marko Laakso (Marko.Laakso@Helsinki.FI), Mikko Kivelä (bolozna@gmail.com)
Issue tracker View/Report issues
Requires R ; mixtools (R-package)
Source files component.xml ExprMixtureModel.r
Usage Example with default values

Inputs

Name Type Mandatory Description
expr Matrix Mandatory Expression matrix
annot AnnotationTable Optional Gene names and other related annotations
groups SetList Optional Sample groups

Outputs

Name Type Description
stats AnnotationTable Distribution statistics for all input genes
report Latex Distribution plots for the selected genes
samplesets Matrix Samples associated with the model components. Provides the posterior probabilities for each sample to belong to the first components.

Parameters

Name Type Default Description
annotCols string "" A comma separated list of gene annotation column names to be shown
group1 string "" ID of the first sample group or an empty string for the first group within the groups input
group2 string "" ID of the second sample group or an empty string for the second group within the groups input
maxLoglik float -800 Maximum of the log likelihood for the mixture model
maxOverlap float 1.00 Maximum overlap between the distribution components
minDifference float 2.0 The difference of the distribution component means is at least this many times the standard deviation of the less variatiating distribution.
minInfo float 0.0 Minimum for the normalized mutual information score when groups 1 and 2 are defined.
minProportion float 0.2 The smaller distribution component should cover at least this proportion of the samples.
sdPoint float 2.0 Cut-off points are visualized at (+,-)sdPoint*mean.
sectionTitle string "Gene expression mixture model" If non-empty, a declaration of a new section with the given name is inserted to the beginning of the combined document.
sectionType string "subsection" Type of LaTeX section: usually one of section, subsection or subsubsection. No section statement is written if sectionTitle is empty.
showIfFail boolean true Show distributions where the model fit algorithm has failed

Test cases

Test case Parameters IN
expr
IN
annot
IN
groups
OUT
stats
OUT
report
OUT
samplesets
case1 properties expr annot groups stats report samplesets

annotCols = name,description,
group1 = case,
group2 = control

case2 properties expr (missing) (missing) stats report samplesets

showIfFail = false,
maxLoglik = 0.0


Generated 2018-12-16 07:42:09 by Anduril 2.0.0