Section 04 · The Dossier E-19
Read-across
When measured data is absent, a registrant may argue that a structurally similar substance fills the gap — but the argument must be made, not assumed.
What the technique does and what it must show
A registration dossier rarely contains a complete set of measured values for every endpoint required at a given tonnage band. Some studies are ethically or technically impractical to repeat; others were simply never conducted for the substance in question. Read-across addresses this by borrowing data from a source substance — one for which measurements exist — and applying them to a target substance on the grounds of structural, metabolic or physicochemical similarity.
The logic is straightforward in principle: compounds that share a common functional group, a common metabolic pathway, or a predictable trend within a homologous series can reasonably be expected to behave alike in certain respects. What regulators scrutinise is whether the registrant has made that case rigorously, or merely assumed it.
ECHA's read-across assessment framework — RAAF, first published in 2015 — sets out the structure that a justification must follow. The registrant must identify the source substance and explain why it is analogous; must document the data being borrowed; must describe the structural relationship in mechanistic terms; and must account for any properties that differ between source and target in ways that could affect the endpoint at issue. A statement that two substances are "similar" is not an argument. The differences matter as much as the similarities, and an honest assessment names both.
The scrutiny read-across attracts during dossier evaluation is considerable. Assessors look for internal consistency: if one endpoint borrows data from substance A, a different endpoint cannot borrow from substance B unless the registrant explains why different analogues are appropriate for different properties. Inconsistency of that kind signals that the analogy was chosen to fit the answer rather than to reflect chemistry.
Where the read-across is poorly supported, ECHA may issue a decision requiring the registrant to generate the missing measured data directly. The technique is a legitimate tool, but it functions as a structured scientific argument, not a shortcut — and the dossier record must show the reasoning in full, not just the conclusion.
Filed inSection 04 — The Dossier