ATP-sensitive potassium (K_ATP) channels control electrical signalling in diverse cell types by coupling cellular metabolism to potassium movement across cell membranes. Their activity is regulated by intracellular adenine nucleotides, with ATP having an inhibitory effect and MgADP having a stimulatory effect on channel activity. More recently, the state-dependency of ATP inhibition has become a matter of controversy; in particular it is unclear whether ATP interacts with the open state of the channel to induce pore closure (e.g. Enkvetchakul et al. 2001; Li et al. 2003). In order to resolve this controversy, we have simulated single-channel activity with the QuB program (Qin & Auerbach, Buffalo University, USA) using a wide range of kinetic models that incorporate both the tetrameric structure of the channel and different topologies of intraburst/interburst transitions. We next performed single-channel and macroscopic analysis of in silico single channel records to obtain ATP dependencies of dwell times and dose-response curves of K_ATP channel inhibition. Our results demonstrated that models which assume that stabilisation of the closed states is the sole mechanism of ATP inhibition are inconsistent with single-channel kinetics of K_ATP channels. We further extended our models to simulate the effects of mutations in the Kir6.2 subunit that affect gating of the unliganded channel. These results are used to discuss the possible contribution of gating and transduction mechanisms to the altered ATP sensitivity observed in these mutants.