Module bioinformatics 

SIMD-Accelerated Bioinformatics Engine.

Production-quality sequence alignment with hardware dispatch:

• Smith-Waterman local alignment (affine gap penalties)
• Needleman-Wunsch global alignment
• BLOSUM62 / nucleotide substitution matrices
• K-mer frequency analysis + MinHash sketching
• FASTA record validation
• Tanimoto metabolite fingerprint similarity

qualia:alignNucleotideSequence → align_nucleotide() qualia:alignProteinSequence → align_protein() qualia:computeKmerFrequency → kmer_frequencies() qualia:validateFastaRecord → validate_fasta_record() qualia:computeMetaboliteSimilarity → tanimoto_similarity()

Structs

AlignmentResult

Full alignment result with traceback and statistics.

AlignmentScore

Lightweight backward-compatible score wrapper.

FastaRecord

GapPenalty

Affine gap penalty model (BLAST defaults: open=-11, extend=-1).

NucleotideMatrix

Nucleotide substitution scores.

PhyloMerge

One agglomeration event in a UPGMA tree: the two child cluster ids merged, the node height, and the id assigned to the new (parent) cluster. Leaves are 0..n; internal nodes are numbered n..2n-1 in merge order, so the last record’s merged_id is the root.

Enums

SequenceAlphabet

Constants

MAX_PHYLO_TAXA

Maximum number of taxa (leaves) a single bounded UPGMA build admits. Bounding the taxon count keeps the working distance matrix on the stack — zero heap.

Functions

align_nucleotide

DNA/RNA Smith-Waterman with BLAST nucleotide defaults.

align_protein

Protein Smith-Waterman with BLOSUM62.

align_sequences

Backward-compatible entry point returning the legacy AlignmentScore.

blosum62_score

BLOSUM62 score for two amino acid bytes. Unknown residues → -4.

build_upgma_tree

Build a UPGMA (Unweighted Pair Group Method with Arithmetic Mean) phylogenetic tree from a flattened n×n symmetric distance matrix (distances[i*n + j]).

calculate_isoelectric_point

Estimates the Isoelectric Point (pI) of a protein sequence using the Henderson-Hasselbalch equation and basic pKa values.

dice_similarity

Dice coefficient between two binary fingerprints.

jaccard_similarity

Jaccard similarity (0.0–1.0) between two MinHash sketches.

kmer_frequencies

Counts all k-mer occurrences; returns (kmer_fnv1a_hash, count) sorted by hash.

minhash_sketch

MinHash sketch: the sketch_size smallest k-mer hashes.

needleman_wunsch

Needleman-Wunsch global alignment with linear gap penalty.

predict_peptide_cleavage

Predicts cleavage sites for Trypsin (cleaves after K or R, unless followed by P). Writes the indices of cleavage into the caller-provided out_indices buffer. Returns the number of cleavage sites found.

smith_waterman

Smith-Waterman local alignment with affine gap penalties. Allocates O(m×n) DP tables — caller should respect MAX_SEQ_LEN.

tanimoto_similarity

Tanimoto (Jaccard) coefficient between two Morgan fingerprints encoded as u64 bitmasks. Multiple words can represent a full extended fingerprint.

translate_dna_to_protein

Translates a DNA sequence into an amino acid sequence using the standard genetic code. Writes directly into the caller-provided out buffer to avoid allocation. Returns the number of amino acids written.

validate_fasta_record

Validates and classifies a FASTA record.