Description Usage Arguments Details
Metrics include overlap between observed and expected signatures, JI between observed and expected signatures, and Fisher p value for selection of observed and expected signatures from a background set.
1 2 | E3CompareSignatures(sampleGroups, group.signature = NULL,
signature.ref = NULL, signature.expected = NULL, numgenes = 5000)
|
sampleGroups |
- named list of character vectors. Each vector should represent a set of samples making up a group. |
group.signature |
- named list of character vectors. Each vector should represent the gene set (signature) associated with a group |
signature.ref |
- named character vector. Each element should link a sample name to a reference group, e.g. KO_ACTA_R1 would be linked to group KO_None |
signature.expected |
- named character vector. Each element should a sample to an expected signature group, e.g. KO_ACTA_R1 would be linked to group WT_ACTA |
numgenes |
- number of genes to use as background in calculation of Fisher metric. |
Warning: the Fisher p value is computed very crudely using a background set of fixed size. This is in principle flawed, but the whole idea of using a Fisher test in such a scenario is flawed because not all genes are equally powered to be part of signatures. Bottom line is that the Fisher p values should be interpreted with care, they may be off even by two/three orders of magnitude.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.