nx_peptide_ext.nx

nx_peptide_ext.nx -- CHEMISTRY SUITE / PEPTIDE EXTENSION rung. Closes the three gaps nx_peptide and nx_supplement_screen declared in their own headers: non-proteinogenic residues, disulfide bridges, and lactam cyclisation. Together those cover the peptides that actually turn up in supplement and research-peptide channels -- which are overwhelmingly modified, cyclic, or both. WHY A BUILDER AND NOT A LONGER ALPHABET. A one-letter code has 26 slots and 20 are spoken for; overloading B/J/O/U/X/Z with Aib/Nle/Orn would be arbitrary, unreadable, and would silently reinterpret any sequence that legitimately used them as ambiguity codes. So a peptide is COMPOSED: runs of standard sequence, explicit non-standard residues by id, then terminal modifications and bridges. Nothing is guessed from a letter. THE VALIDATION IS THE POINT. Oxytocin and vasopressin each land on their published monoisotopic mass ONLY if the C-terminal amide AND the disulfide are both applied -- miss the amide and you are 0.98 Da out, miss the bridge and you are 2.02 Da out. Both are checked against literature (1006.4367 and 1083.4380), and the gate ALSO runs the no-bridge and no-amide variants as negative controls, so the test cannot pass by accident. D-amino acids need no entry here: a stereoisomer has identical mass. GHRP-6's D-Trp and D-Phe, ipamorelin's D-2-Nal -- all already correct in nx_peptide. Chirality matters to the receptor, not to the balance. All INTEGER, _q4 = x10^4 Da, matching nx_peptide. Grounding (cited; researcher-groundable): iupac_nonproteinogenic_residue_formulas (extended residue masses) disulfide_bond_two_hydrogen_loss (-2.01565 Da per S-S) lactam_cyclisation_water_loss (-18.01056 Da per bridge) published_monoisotopic_masses_oxytocin_vasopressin (validation anchors) genealogy_id: peptide_chemistry + nishi_chem_suite