Check compatibility of seed ID pairs in input rows — seed_id_compatibility

Creates a long-format permutation table where each row represents one unique unordered pair of seed IDs from the same input row, then flags whether each pair is compatible via direct or secondary graph connections.

Usage

seed_id_compatibility_check(
  data,
  id_types = c("HMDB", "KEGG", "CHEBI", "PUBCHEM"),
  delimiter = c(";", ","),
  verbose = FALSE,
  edge_table = NULL,
  handle_partially_compatible = FALSE,
  handle_completely_incompatible = FALSE,
  completely_incompatible_priority = c("HMDB", "CHEBI", "PUBCHEM", "KEGG")
)

Arguments

data: Data frame with zero or more of the columns HMDB, KEGG, CHEBI, and PUBCHEM. Column names are matched case-insensitively against these exact names.
id_types: Character vector of ID types to use. Choose from HMDB, KEGG, CHEBI, and PUBCHEM.
delimiter: Character string indicating whether multiple IDs within one cell are separated by semicolons or commas. Accepted values are ";", ",", "semicolon", or "comma".
verbose: Logical; if TRUE, prints pairwise mapping and edge construction diagnostics to the console.
edge_table: Optional precomputed bidirectional edge table with columns id1, type1, id2, type2. If NULL, the table is built internally.
handle_partially_compatible: Logical; if TRUE, partially compatible features are cleaned by retaining only IDs from compatible pairs.
handle_completely_incompatible: Logical; if TRUE, completely incompatible features are cleaned by retaining a single ID according to completely_incompatible_priority.
completely_incompatible_priority: Character vector defining the namespace priority for resolving completely incompatible features. Supported values are HMDB, KEGG, CHEBI, and PUBCHEM. The default priority is c("HMDB", "CHEBI", "PUBCHEM", "KEGG").

Value

Named list with at least five data frames:

ID_pair_compatibility: Long-format table with one unique unordered seed-ID pair per input row. The first column original_row_id stores the original input row name. The table also includes pair_compatible, compatibility_path (direct, secondary, no_match), and grouped all_seed_ids_compatible.
data_with_compatibility: Original input data with appended all_seed_ids_compatible per input row (rows with fewer than two seed IDs are TRUE).
feature_compatibility_summary: One row per input feature summarizing compatibility counts and assigning fully_compatible, partially_compatible, or completely_incompatible.
data_after_handling: Feature-level table after optional automatic handling. If no handling is enabled, this matches the raw feature-level output aside from added summary columns.
ID_pair_compatibility_after_handling: Pair-level compatibility table recomputed from data_after_handling.

If either handling option is enabled, the return object also includes handling_summary_text and handling_summary_metrics.