In this work we use for the first time an unconventional and bio-inspired model of computation known as a virus machine (in short, VM) to model workflow patterns. VMs are novel computing paradigms inspired by biological viruses and their transmissions. VMs are capable of building sequential computing algorithms (they are Turing-universal). In this work, a young variant of VMs, called parallel VMs (in short, PVMs) is used for defining and analysing workflow patterns. More precisely, fundamental workflow patterns, such as sequences and various types of parallel splits and joins, are defined using PVMs. Besides the mathematical formulation of PVMs, their graphical representations have attractive aspects for workflow patterns. Such patterns can form the basis for modelling tasks, consumption of resources, to support for instance in decision making.