Up: Component summary Component

EdgeR

Computes the digital expression fold change and p-value based on counts. The counts are normalized before fold change is calculated. Only the unpaired samples test is implemented.

Version 1.0
Bundle sequencing
Categories
Authors Riku Louhimo (Riku.Louhimo@Helsinki.FI)
Issue tracker View/Report issues
Requires edgeR (R-bioconductor)
Source files component.xml EdgeR.r
Usage Example with default values

Inputs

Name Type Mandatory Description
counts CSV Mandatory A matrix of read counts for gene, transcript or other tracked genomic feature (rows) in the sample set (columns). Rows with missing values are ignored.
groups SampleGroupTable Mandatory Sample groups for counts input columns.
reads CSV Optional Total number of reads for each sample in counts. If reads is not set, the sum of counts for each sample is used instead.

Outputs

Name Type Description
logratio CSV Logratios computed based on target and reference groups. The overall concentration of a tag, p-value for logratios and, if correction was used, the corrected p-values are also included for each row in the input.
plot Latex MA plot of pvalue dispersion with the top 500 genes highlighted.
normalized CSV Normalized counts per tag.

Parameters

Name Type Default Description
correction string "fdr" Method for multiple hypothesis correction of exact test p-values. One of 'none' (no correction), 'fdr' (default), 'robustfdr', 'BY', 'holm' or 'bonferroni'.
dispersion string "tag" Sets dispersion method for exact test. One of 'tag' (tagwise) or 'common'.
normMethod string "TMM" Method for normalization of counts. One of 'TMM', 'RLE' or 'upperquartile'.
normQuartile float 0.75 Quartile for scaling factor in quartile normalization. Only used when normMethod='quartile'.
onlyNorm boolean false Only normalize counts.
pLimitPlot float 0.05 P-Value threshold for defining a differentially expressed gene. Only used for dispersion plot output.
referenceGroup string "" Name of reference group in groups input. Only used if normMethod='TMM'. If referenceGroup is empty, the sample with an upper quartile closest to the mean upper quartile is used as reference.
targetGroup string (no default) Name of target (case) group in groups input. Only used if normMethod='TMM'.

Test cases

Test case Parameters IN
counts
IN
groups
IN
reads
OUT
logratio
OUT
plot
OUT
normalized
case1 properties counts groups reads logratio plot (missing)

referenceGroup = G2,
targetGroup = G1,
correction = fdr

case2 properties counts groups (missing) logratio (missing) (missing)

referenceGroup = G2,
targetGroup = G1,
dispersion = common,
normMethod = upperquartile,
normQuartile = 0.7

case3_emptyCounts properties counts groups reads (expecting failure) (expecting failure) (expecting failure)

targetGroup = G1

case4_onecol properties counts groups (missing) logratio (missing) normalized

referenceGroup = G2,
targetGroup = G1,
correction = BY,
pLimitPlot = 1

case5_onlyCounts properties counts groups reads (missing) (missing) normalized

referenceGroup = G2,
targetGroup = G1,
onlyNorm = true


Generated 2018-12-17 07:42:21 by Anduril 2.0.0