Skip to contents

Metabolomics pathway enrichment

Source: R/pathway_enrichment.R
run_fella.Rd

Run FELLA for metabolomics enrichment in KEGG pathways. Return results in a format similar to gprofiler2::gost() results.

Usage

run_fella(input, background, fella.data, method = "hypergeom", niter = 1e+05)

Arguments

input

character vector, input features (KEGG IDs)

background

character vector, background features (KEGG IDs)

fella.data

FELLA database returned by FELLA::loadKEGGdata()

method

character, FELLA enrichment method, one of "hypergeom", "diffusion"

niter

integer, number of iterations with which to estimate p-values by simulation. Only applies for the diffusion method of enrichment.

Value

data table of pathway enrichment results:

term_size

double, number of KEGG IDs that are annotated to the term

query_size

integer, number of KEGG IDs that were included in the query

intersection_size

double, the number of KEGG IDs in the input query that are annotated to the corresponding term

term_id

character, unique term/pathway identifier

source

character, database source of term/pathway

computed_p_value

double, nominal enrichment p-value

kegg_id

character, KEGG ID for term_id

Examples

if (FALSE) {
# Make KEGG database
kegg_db = "~/KEGGdb/test"
make_kegg_db(kegg_db)

# Get FELLA data 
fella.data = FELLA::loadKEGGdata(databaseDir = kegg_db, 
                                 internalDir = FALSE,
                                 loadMatrix = c("hypergeom","diffusion"))

# Get input features
# Metabolites in SKM-GN:8w_F1_M1 that map to a KEGG ID
input = c("C02918","C00195","C01967","C00016","C04438","C02294","C00003","C00006",
          "C00157","C00350","C04475","C00344","C06125","C00550","C00387","C04230",
          "C00073","C00864","C00670","C00836","C00319")

# Get universe/background list 
background = MotrpacRatTraining6moData::GENE_UNIVERSES$gene_symbol$METAB$`SKM-GN`

# Example 1: method "hypergeom"
res = run_fella(input, background, fella.data, method="hypergeom")

# Example 2: method "diffusion" (more powerful but slower and more difficult to interpret)
res = run_fella(input, background, fella.data, method="diffusion")
}