Danny2045/pfdhfr-caller

Sovereign, offline pyrimethamine-resistance caller for P. falciparum dhfr. Validated on real 3D7 and 20 East African field isolates. # pfdhfr-caller Call the classic pyrimethamine-resistance mutations in *Plasmodium falciparum* dihydrofolate reductase (dhfr) from an in-frame coding sequence. Given a FASTA of a dhfr coding sequence (CDS), the tool reads codons 51, 59, 108 and 164, translates each, compares it to the wild-type (3D7) residue, and prints the per-codon call plus a single combined resistance category. It is deliberately small. For an explicit statement of what it does and does not cover, see `docs/SCOPE.md`. ## What it checks | Codon | Wild-type (3D7) | Canonical resistance mutation | |------:|:---------------:|:------------------------------| | 51 | N (Asn) | N51I | | 59 | C (Cys) | C59R | | 108 | S (Ser) | S108N | | 164 | I (Ile) | I164L | S108N is treated as the pivotal mutation; N51I, C59R and I164L stack on top of it to raise the level. Any residue that differs from the wild-type counts as a mutation and is labelled with the amino acid actually observed — not only the canonical resistance allele. ## Combined categories it can emit - `sensitive (wild-type at all four codons)` - `pyrimethamine resistance` — S108 mutated, fewer than three codons mutated - `high-level pyrimethamine resistance` — S108 mutated and at least three of the four codons mutated - `atypical genotype (resistance codons mutated but S108 wild-type)` - `indeterminate (one or more codons could not be read)` The category is a coarse, rule-based summary of the four-codon genotype, not a clinical recommendation. See `docs/SCOPE.md`. ## Run it From the repository root: ``` python3 src/call_pfdhfr.py ``` Run the tests (plain asserts, no framework required): ``` python3 tests/test_caller.py ``` ## Example: real 3D7 reference ``` $ python3 src/call_pfdhfr.py data/3d7_dhfrts_real.fasta >lcl|XM_001351443.1_cds_XP_001351479.1_1 (1827 bp) codon 51: AAT -> N N51 codon …