Quarter in progress
A provisional report for the current quarter. Figures will change as records arrive; the standing observations will not.
Standing observations
Six things have been true of this dataset in every quarter since it opened, and none of them has improved materially.
- Purity is over-measured and content is under-measured. Purity appears on nearly every record; peptide content on roughly half; counter-ion on fewer. The measurement that determines what you actually receive is the one least often made.
- Endotoxin is absent from three quarters of records. It is the only property no chromatographic result bears on and it is the one most often omitted.
- Methods are not stated. Gradient duration, reporting threshold, integration convention and reference standard remain absent from most supplier certificates.
- Under-fill outnumbers over-fill by about three to one in the gravimetric subset. That asymmetry is difficult to explain as random fill variation.
- Round-number declarations are more optimistic than precise ones, by roughly a factor of two in divergence from independent measurement.
- Supplier-supplied samples read high. Small on average, occasionally large, and entirely avoidable by buying the material like anyone else.
| Rule | Flagged this quarter | Rate | Severity |
|---|---|---|---|
| AX-R01 — Mass accuracy outside class limit | 12 | 2.64% | HIGH |
| AX-R02 — Declared impurity arithmetic does not close | 0 | 0.00% | MEDIUM |
| AX-R03 — Declared largest single impurity exceeds declared total | 0 | 0.00% | HIGH |
| AX-R04 — Digit preference in declared purity | 44 | 9.67% | LOW |
| AX-R05 — COA purity diverges from independent median | 0 | 0.00% | HIGH |
| AX-R06 — Recovered mass outside label claim tolerance | 15 | 3.30% | MEDIUM |
| AX-R07 — Declared gravimetric mass balance does not close | 3 | 0.66% | MEDIUM |
| AX-R08 — Endotoxin above index specification | 11 | 2.42% | HIGH |
| AX-R09 — Impurity fingerprint shared across suppliers | 31 | 6.81% | LOW |
| AX-R10 — Supplier-supplied sample reads high against buyer-sourced | 7 | 1.54% | MEDIUM |
| AX-R11 — Retention time outside compound window | 15 | 3.30% | MEDIUM |
| AX-R12 — Peak asymmetry above the integration limit | 1 | 0.22% | MEDIUM |
What would improve the index most
- Named impurities instead of totals. One laboratory does this and it is the source of nearly every cross-record finding the index has produced.
- Content, water and counter-ion as a default panel. Three additional measurements would convert most of this index from a purity register into a mass register.
- Submissions on the thin cells. Several supplier–compound cells have n=1 or n=2 and no amount of analysis fixes that. The only remedy is a vial and a courier.
- Buyer-sourced material only. The cleanest thing any submitter can do is buy the vial like a customer.
The index publishes its whole dataset as CSV and JSON under CC BY 4.0 precisely so that anyone who disagrees with the analysis above can redo it. That is the only durable answer to the objection that an index is only as good as its own judgement.
Appendix
Quarter in numbers
Records submitted
455
n=455
Batches
75
n=75
Median purity
98.93%
n=455
IQR
1.12pp
n=455
Suppliers represented
26
n=455
Compounds represented
34
n=455
Records flagged
131
n=455
Buyer-sourced share
86.4%
n=455