Further advances in pharmacogenetic treatment and pharmacogenomic research will depend on large international collaborations and 'biobanks'.

One of the biggest sticking points in pharmacogenomic research is size - studies of very large groups of people are needed to ensure that results are statistically significant.

This can be a big challenge. Large numbers of patients must be collected and data on their condition captured in the same way. For many conditions, diagnosis might be difficult, or a disease may vary and have different underlying causes (as with schizophrenia).

If a side-effect is serious but rare, collecting enough patients can be extremely difficult.

Studies are therefore increasingly international in size, with researchers from different countries collaborating and pooling results. The organisation of different groups, operating under different regulations and often with slightly different techniques presents considerable practical challenges.

A related approach is to make use of 'biobanks' or 'cohort studies', which collect and store health information and biological samples from large numbers of people. The UK Biobank is recruiting 500 000 people aged 40-69. Similar studies have been established in other countries.

These biobanks have the advantage that large numbers are people are providing information to researchers. The disadvantage is that volunteers are largely a random cross-section of the population, so the numbers of people with some less common conditions will be small.

There are also initiatives to coordinate the collection of data on adverse reactions once medications are being used. In late 2007 a not-for-profit body - the Serious Adverse Events Consortium - was set up in the USA to collect and share data on adverse reactions across the world, starting with drugs which appear to be toxic to some people's livers and serious ones which are known to sometimes cause hypersensitivity reactions.