Constructs hypergraph from pairs data by binning, filtering, and building sparse incidence matrix.
Usage
build(x, ...)
# S4 method for class 'MultiWayContacts'
build(x, bin_size = 1000000L, quantile = 0.85, min_multiway = 2)Arguments
- x
MultiWayContacts object with loaded pairs.
- ...
Additional arguments (not used).
- bin_size
Integer. Bin size in base pairs (default: 1000000L).
- quantile
Numeric (0-1). Quantile threshold for filtering bin pairs by contact frequency (default: 0.85).
- min_multiway
Integer. Minimum contacts per hyperedge to retain (default: 2).
Value
MultiWayContacts with populated hypergraph slots: hypergraph (incidence matrix), weights (normalizations), bin_info, hyperedge_reads, multiways, and bin_size.
Details
Steps: bin positions → filter by quantile → build incidence matrix → deduplicate patterns → filter by min_multiway.
Weight normalizations: raw (frequency), log (log-transformed), by_order (normalized within order), minmax (scaled to [0,1]).
Higher quantile = stricter filtering. Typical: 0.8-0.95.
Examples
if (FALSE) { # \dontrun{
# Default parameters
mc <- MultiWayContacts("sample.pairs.gz") |> import() |> build()
# Custom parameters
mc <- mc |> build(bin_size = 500000L, quantile = 0.9, min_multiway = 3)
} # }