Biotransformation of chemical contaminants is of importance in various natural and engineered systems. However, in complex microbial communities and with chemical contaminants at low concentrations, our current understanding of biotransformation at the level of enzyme–chemical interactions is limited. Here, we explored an approach to identify associations between micropollutant biotransformation and specific gene products in complex microbial communities, using association mining between chemical and metatranscriptomic data obtained from experiments with activated sludge grown at different solid retention times. We successfully demonstrate proportional relationships between the measured rate constants and associated gene transcripts for nitrification as a major community function, but also for the biotransformation of two nitrile-containing micropollutants (bromoxynil and acetamiprid) and transcripts of nitrile hydratases, a class of enzymes that we experimentally confirmed to produce the detected amide transformation products. As these results suggest that metatranscriptomic information can indeed be quantitatively correlated with low abundant community functions such as micropollutant biotransformation in complex microbial communities, we proceeded to explore the potential of association mining to highlight enzymes likely involved in catalyzing less well-understood micropollutant biotransformation reactions. Specifically, we use the cases of nitrile hydration and oxidative biotransformation reactions to show that the consideration of additional experimental evidence (such as information on biotransformation pathways) increases the likelihood of detecting plausible novel enzyme–chemical relationships. Finally, we identify a cluster of mono- and dioxygenase fourth-level enzyme classes that most strongly correlate with oxidative micropollutant biotransformation reactions in activated sludge.