matched molecular pair
Imagine running an experiment where you change exactly one thing and keep everything else identical — the cleanest possible way to learn what that one thing does. A matched molecular pair is two molecules that are the same everywhere except for a single, well-defined change at one site, such as swapping a hydrogen for a fluorine or a methyl for an ethyl.
Formally, a matched molecular pair is a pair of compounds related by a single, localized chemical transformation, with the rest of the structure held constant. The difference in any measured property between the two — potency, solubility, metabolic stability, permeability — is the contribution attributable to that one transformation. Collecting thousands of such pairs across many projects lets matched molecular pair analysis estimate the average effect of common substituent swaps, turning anecdotes into statistics.
MMPs are powerful because they isolate cause from a noisy background, and they scale to whole corporate databases. The honest caveat is context dependence: the average effect of, say, adding a fluorine is just an average. In any specific molecule and pocket the same swap can do the opposite, so MMP rules guide ideas but do not replace making and testing the actual compound.
Compound A bears a hydrogen where compound B bears a chlorine, and B is ten-fold more potent — that single hydrogen-to-chlorine pair quantifies what the chlorine buys at that position.
One controlled change, one measured difference — the basic unit of quantitative SAR.
An activity cliff is a special, dramatic kind of matched molecular pair: a single small change that produces a surprisingly large activity difference.