Struct Haplotype 

pub struct Haplotype { /* private fields */ }

Sparse representation of one haplotype — a reference overlay of variants.

Instead of materializing a full haplotype sequence (which would require ~250MB per haplotype for human chr1), this stores only the differences from the reference as a sorted set of variants in a COITree for efficient range queries.

Fragment extraction works by walking the reference sequence and substituting alt alleles at variant positions on the fly.

Because coitrees requires Copy + Default for metadata, we store variant indices (as u32) in the tree and keep the actual variant data in a separate Vec.

Implementations

impl Haplotype

pub fn allele_index(&self) -> usize

Return the allele index of this haplotype.

pub fn extract_fragment( &self, reference: &[u8], ref_start: u32, fragment_len: usize, ) -> (Vec<u8>, Vec<u32>, u32)

Extract a fragment from this haplotype at the given reference coordinates.

Walks the reference from ref_start and produces fragment_len bases, substituting alternate alleles where this haplotype has variants. Returns the fragment bases, a list of reference positions corresponding to each fragment base (for golden BAM coordinate mapping), and the haplotype-coordinate position the fragment starts at (for per-haplotype methylation lookups).

Arguments

• reference — Full reference sequence for this contig.
• ref_start — 0-based start position on the reference.
• fragment_len — Desired number of output bases.

Returns

A tuple of (fragment_bases, ref_positions, hap_start):

• fragment_bases[i] is the base at fragment-internal position i.
• ref_positions[i] is the reference position corresponding to fragment_bases[i]. For inserted bases, the reference position is that of the base preceding the insertion.
• hap_start is the first haplotype-coordinate position covered by the fragment. Equals ref_start plus net upstream insertion length (insertions add bases, deletions remove bases). With no upstream indels it equals ref_start.

pub fn hap_position_for(&self, ref_pos: u32) -> u32

Map a 0-based reference position to its 0-based haplotype position.

Sums net length changes (alt_len - ref_len) across every variant strictly upstream of ref_pos (i.e. var.ref_pos + var.ref_len <= ref_pos). Variants that straddle ref_pos do not yet contribute — the caller is presumed to be at a position outside any deletion’s interior (otherwise the haplotype position would be ambiguous).

Used by Self::extract_fragment to compute hap_start and by methylation-table construction to relate haplotype scans to per- haplotype indices. Implementation is O(log n) via binary search over the precomputed Self::end_prefix_deltas index.