Related-substances profiling

A total-impurities figure tells you how much is wrong. A named related-substances table tells you what is wrong, and that is a categorically more useful piece of information — because impurity identity distinguishes a synthesis problem from a storage problem, and because the vector of named impurities and their levels is a fingerprint of one particular manufacturing run.

Assignment workflow

1. Integrate every peak above the reporting threshold (0.05 % area for the index) and record retention time, area, area percent and, where a PDA is available, the UV spectrum at peak apex.
2. Acquire mass spectra across each impurity peak and compute the mass difference from the target.
3. Match the mass difference against the expected impurity set for the compound: deletion sequences (−57 Gly, −71 Ala, −87 Ser, −113 Leu/Ile, −128 Gln/Lys, −137 His, −163 Tyr), modifications (+16 oxidation, +42 acetylation, +96 TFA adduct, −18 aspartimide or cyclisation, +1 deamidation, +2 disulfide reduction), and protection remnants (+56 tBu, +222 Fmoc).
4. Where the mass difference is zero, the impurity is an isomer — a diastereomer, a regioisomer, or a β-Asp rearrangement product — and mass spectrometry contributes nothing further. Assignment then rests on retention order and, if available, co-injection with an authentic standard.
5. Confirm the aromatic content from the 280/214 nm absorbance ratio: a peak whose ratio matches the parent contains the same aromatic residues, which excludes a large class of misassignments.

Common mass differences and what they mean

Fingerprint matching

Two independent syntheses of the same compound do not produce the same four impurities at the same four levels. The index matches on the named-impurity vector — identical identities, identical rank order, each level agreeing within 0.05 % area or 10 % relative, whichever is larger, with a minimum of four named impurities. A match indicates one bulk under two labels.

Twenty-nine such matches exist across 19 distinct bulks. Nothing about that is improper — contract synthesis is how the industry works — but it means two suppliers' published results on matched batches are one measurement and not two, and the index counts them once in aggregates and cross-links both batch pages. Rule AX-R09 documents the predicate.