Agent-based modelling of iron cycling bacteria provides a framework for testing alternative environmental conditions and modes of action.

Iron-reducing and iron-oxidizing bacteria are of interest in a variety of environmental and industrial applications. Such bacteria often co-occur at oxic-anoxic gradients in aquatic and terrestrial habitats. In this paper, we present the first computational agent-based model of microbial iron cycling, between the anaerobic ferric iron (Fe 3+ )-reducing bacteria Shewanella spp. and the microaerophilic ferrous iron (Fe 2+ )-oxidizing bacteria Sideroxydans spp. By including the key processes of reduction/oxidation, movement, adhesion, Fe 2+ -equilibration and nanoparticle formation, we derive a core model which enables hypothesis testing and prediction for different environmental conditions including temporal cycles of oxic and anoxic conditions. We compared (i) combinations of different Fe 3+ -reducing/Fe 2+ -oxidizing modes of action of the bacteria and (ii) system behaviour for different pH values. We predicted that the beneficial effect of a high number of iron-nanoparticles on the total Fe 3+ reduction rate of the system is not only due to the faster reduction of these iron-nanoparticles, but also to the nanoparticles' additional capacity to bind Fe 2+ on their surfaces. Efficient iron-nanoparticle reduction is confined to pH around 6, being twice as high than at pH 7, whereas at pH 5 negligible reduction takes place. Furthermore, in accordance with experimental evidence our model showed that shorter oxic/anoxic periods exhibit a faster increase of total Fe 3+ reduction rate than longer periods.