Influenza A virus poses one of the greatest infectious disease challenges of the 21st Century. It is a ubiquitous avian pathogen with vast antigenic diversity that hinders conventional vaccine approaches, especially in low-value livestock species like poultry. It causes huge economic losses and drains public health budgets. Surveillance programmes generate huge amounts of viral sequence data; surpassing 1 million entries on Genbank. Some aspects of virus behaviour can be predicted from these sequences, but many important facets cannot; this wealth of data therefore represents an underutilised resource. We think that advances in computational approaches mean that the construction of modelling tools with genuine predictive power for the future evolution and spread of avian influenza is possible. To achieve this, we have assembled an international team of experts with interdisciplinary expertise in mathematical modelling, influenza, and the infectious disease-public and animal health interface. Importantly this includes Chinese colleagues who run a surveillance programme in the epicentre of viral diversity. The prediction tool will be the sum of three separate models: one which identifies key viral sequence polymorphisms; one which simulates virus evolution within host under selection pressure; and one that integrate outputs from the first two along with additional inputs from surveillance programs. The primary data inputs are virus sequence information, both at quasi-species and consensus level. We will parameterise the models from existing data (public and unpublished data held by the team) and a series of planned "wet lab" experiments that measure virus fitness. We wish the tool to be of use to stakeholders such as the OIE and WHO as well as small and large poultry holders; development of it will therefore be informed by a series of data collection exercises to get input from these groups of people as to what they require from the scientists.