PAINS
Some compounds are the screening equivalent of a person who says 'yes' to every question — they show up as hits against almost any target, in almost any assay, which is exactly why you should distrust them. PAINS, short for pan-assay interference compounds, are chemical classes notorious for producing misleading positive signals not through genuine, selective binding but through troublesome chemistry.
These molecules misbehave by various means: some are chemically reactive and covalently modify proteins indiscriminately, some are redox-active and generate reactive oxygen, some chelate metals, fluoresce, or aggregate. The common feature is promiscuity — a real, specific drug hits a few related targets, whereas a PAINS hits everything, which is a hallmark of artifact rather than mechanism.
Medicinal chemists use published substructure filters (such as the Baell–Holloway PAINS alerts) to flag these motifs in screening libraries and hit lists, so they can be scrutinized or set aside. The flag is a warning, not a verdict: a few legitimate drugs contain PAINS-like substructures, so a flagged compound should be tested with orthogonal, artifact-resistant assays rather than automatically discarded.
A hit list is filtered for PAINS substructures; several flagged catechols and rhodanines are deprioritized after they also score across unrelated counter-screens.
Substructure flags plus promiscuous behavior across assays mark a likely PAINS artifact.
PAINS filters are guidance, not gospel — they over-flag and miss some artifacts. The reliable test is whether a hit behaves like a real, selective binder across orthogonal assays, not merely whether it trips a substructure alert.