Enrichment analysis on offspring terms by permutation test

dag_enrich_on_offsprings_by_permutation(
  dag,
  value,
  perm = 1000,
  min_offspring = 10,
  verbose = simona_opt$verbose
)

Arguments

dag: An ontology_DAG object.
value: A numeric value. The value should correspond to terms in dag@terms.
perm: Number of permutations.
min_offspring: Minimal size of the offspring set.
verbose: Whether to print messages.

Value

A data frame with the following columns:

term: Term names.
stats: The statistics of terms.
n_offspring: Number of offspring terms of t (including t itself).
log2_fold_enrichment: defined as log2(s/mean) where mean is calculated from random permutation.
z_score: Defined as (s - mean)/sd where mean and sd are calculated from random permutation.
p_value: P-values from permutation test.
p_adjust: Adjusted p-values from the BH method.

The number of rows in the data frame is the same as the number of terms in the DAG.

Details

In the function dag_enrich_on_offsprings(), the statistic for testing is the number of terms in each category. Here this funtion makes the testing procedure more general

The function tests whether a term t's offspring terms have an over-represented pattern on values in value. Denote T as the set of t's offspring terms plus t itself, and v as the numeric vector of value, we first calculate a score s based on values in T:

s = mean_{terms in T}(v)

To construct a random version of s, we randomly sample n_T terms from the DAG where n_T is the size of set T:

sr_i = mean_{n_T randomly sampled terms}(v)

where index i represents the i^th sampling. If we sample k times, the p-value is calculated as:

p = sum_{i in 1..k}(I(sr_i > s))/k

Examples

# \dontrun{
dag = create_ontology_DAG_from_GO_db() 
#> relations: is_a, part_of
value = runif(dag_n_terms(dag)) # a set of random values
df = dag_enrich_on_offsprings_by_permutation(dag, value)
#> Error in dag_enrich_on_offsprings_by_permutation(dag, value): could not find function "dag_enrich_on_offsprings_by_permutation"
# }
1
#> [1] 1