complexStatus {PCpheno} | R Documentation |
Categorize the complex whether or not a complex is composed of a significant number of genes involved in a particular phenotype than expected by chance.
complexStatus(data, phenotype, interactome, threshold=0.05)
data |
Output from CoHyperG test |
phenotype |
List of gene names inducing an observed phenotype, e.g., list of essential gene names (see package SLGI) |
interactome |
A binary matrix composed of genes (rows) and biological complexes (columns) (see package ScISI) |
threshold |
pvalue threshold (default 0.05) |
We form four distinct categories from A to D to characterize how a complex might be involved in a particular phenotype (according to the number of genes it contains and that are involved in a particular phenotype - see also CoHyperG function)
The returned value is a list with components:
A |
"interesting" complexes, complexes with a significant number of interesting genes, i.e., genes that participate to a particular phenotype (at a given p-values threshold) |
B |
complexes with a NON significant number of interesting genes BUT that SHARE genes with complexes from the A status |
C |
complexes with a NON significant number of interesting genes AND that DON'T SHARE interesting genes with complexes from cat A |
D |
complexes WITHOUT interesting genes, i.e. the one involved in the studied phenotype |
N. LeMeur
data(ScISI) data(essglist) essential <- names(essglist) CoparamsESS <- new("CoHyperGParams", geneIds=essential, universeGeneIds=rownames(ScISI), annotation="org.Sc.sgd.db", categoryName="ScISI", pvalueCutoff=0.01, testDirection="over") sign<- hyperGTest(CoparamsESS) test05 <-complexStatus(data=sign, phenotype=essential, interactome=ScISI, threshold=0.05)