concoct: concoct: DRUID wrapper
In diogocamacho/druid: Drug Indication Discoverer

Description Usage Arguments Value

View source: R/concoct.R

This function is a wrapper for a start to finish to identify drugs/drug profiles that mimic or revert a given profile of interest. For a given gene set of interest, the algorithm will compute the cosine similarity between this vector and the gene sets in a given collection of drug profiles. Additionally, the algorithm will generate a set of random gene sets that will be used to determine the probability of a given profile similarity being random. The output of the wrapper is a data frame with all the (unranked) DRUID scores.

1 2	concoct(dge_matrix, num_random, druid_direction, fold_thr, pvalue_thr, entrez)

`dge_matrix`	This is a 2 column matrix for gene expression changes, where column 1 is the gene fold change and column 2 is the corresponding p-value for the fold change. NOTE: Use log2 of the fold changes as output, for example, from 'limma' or 'DESeq2'.
`num_random`	Number of random sets to be generated to calculate significance of enrichment. Defaults to 1,000.
`druid_direction`	Desired effect for DRUID to run on: "pos" mimics query phenotype, "neg" reverts query phenotype. Defaults to "neg".
`fold_thr`	Threshold for the fold change to be considered. Defaults to 0 (i.e., log2(1), where fold change is not used as filter)
`pvalue_thr`	Threshold for the p-value of the fold change to be considered. Defaults to 0.05.
`entrez`	EntrezIDs for genes in differentially expressed set. Must be same order as the input matrix.
`tfidf_matrix`	tf-idf matrix drug-gene matrix. Column names are Entrez IDs. Computed with `ctfidf`