nx_peptide_formula.nx

nx_peptide_formula.nx -- LIAR-KILLER for the peptide mass stack. Computes a peptide's monoisotopic mass by a SECOND, INDEPENDENT path: counting the C/H/N/O/S atoms of every residue and multiplying by CODATA element masses, rather than summing a pre-rounded residue-mass table (nx_peptide). WHY THIS IS REAL LIAR-KILLING, NOT SELF-CONSISTENCY. nx_peptide sums a table of residue masses. If that table had a typo, its gate -- which checks the sum against a literature number -- would still catch a SINGLE-peptide error, but a systematic table error could hide. This organ derives the same masses from a DIFFERENT primitive (atom counts x element masses), and the gate demands three things agree: (1) this elemental mass == nx_peptide's residue-sum mass (two methods) (2) both == the PUBLISHED monoisotopic mass (external) (3) the derived MOLECULAR FORMULA == the PUBLISHED formula (external) For a table typo to survive, the SAME error would have to appear in the residue-mass table AND the atom-count table AND match the published formula AND the published mass -- which is not a mistake, it is a conspiracy. Two independent representations converging on external ground truth is the definition of a killed liar. The molecular-formula check (3) is the sharpest: angiotensin II is the published C50H71N13O12, bradykinin C50H73N15O11. Those are hand-checkable integers, not a mass that hides rounding. If the atom table is wrong the formula is visibly wrong. All INTEGER, _q4 = x10^4 Da, matching nx_peptide. Grounding (cited; researcher-groundable): codata_monoisotopic_atomic_masses_H_C_N_O_S published_molecular_formulas_angiotensin_ii_bradykinin iupac_residue_elemental_compositions genealogy_id: peptide_chemistry + liar_killer