How we weight what we cite.
Every compound claim in the Pivotal library traces to a source. Not all sources carry equal weight. This page describes the framework we apply before any claim reaches a client document or atlas drawer.
Mechanistic biology
The primary tier. Receptor binding, signaling cascades, cellular and subcellular effects, pharmacokinetics confirmed in controlled settings. This is the anchor for every compound classification. If the mechanism is not established, the compound is not in the library.
Operator evidence
Longitudinal outcome data from structured protocols. Response vectors, lab trajectory, tolerability patterns. The operator layer sits at Tier 02 because it is the most proximate to the actual use context, but it carries observational status rather than controlled-trial weight.
Reciprocal regulatory registries
The regulatory record across jurisdictions: FDA, EMA, WADA, and their functional counterparts in Russia, China, Japan, and Korea. These are treated as descriptive artifacts, not arbiters of validity. A compound banned by WADA or unscheduled by the FDA is a data point about regulatory posture, not a judgment on mechanism or safety. The reciprocal view matters: compounds used clinically in one jurisdiction but restricted in another reveal the limits of any single registry's scope.
Case literature and published observational data
Case series, n-of-1 reports, and published observational studies. Useful for hypothesis generation and for filling gaps where controlled trials are absent. Weighted below Tiers 01 and 02 because confounders cannot be excluded. Included in the library with explicit provenance tagging.
Every citation in the Pivotal atlas and client documents is re-verified against the source file in the evidence corpus before it reaches a drawer or protocol page. Verification is not optional on compounds where the mechanistic literature is contested or where the regulatory posture has changed since the initial library entry.
When a claim is inferred from adjacent evidence rather than direct study, the document carries an explicit INFERRED tag. When a claim is contested across sources, the document carries a CONTESTED tag with both positions noted. No claim is presented as settled when the evidence is not.