The rational building of transcriptional circuits has become a central preoccupation in synthetic biology.However, due to the highly-interconnected nature of biological systems and the limited guidance of in silico modeling, the circuit building is usually in a low-throughput, trial-and-error manner.Here we show that refined biological parts permit parameter-free models for the transcriptional circuits;this allows the parameters from the characterization of individual biological parts to be sufficient to predict the functional outcome when assembled together.As a proof, the parameter sheets of 52 promoters and 34 operator settings were randomly paired to yield 81 NOT gates, all of which proved to be in high precision to the model prediction (R2=0.95).We further combined the parameter sheets with in silico computation to design more than 30,000 incoherent feedforward networks (IFFNs) and genetically implement some of them with typical function;all the circuits functioned in high precision.