To answer this question, the Science Advisory Board, an on-line panel of 17,300 scientists and medical professionals from 62 countries, asked its members to judge whether computational methods of determining the structure of a drug is more effective than the lab-based approach, or vice versa. They were told to consider the financial constraints on research budgets when choosing their answers.

70 per cent of the 1,600 responses favoured a combination approach that would use both computational and analytical methods to tackle structure determination and characterisation.

Computational methods are based on molecular simulation or quantum mechanics calculations. Molecular simulation, in particular, is well suited to examining the structures of large molecules like proteins. In contrast, quantum mechanics allows for highly precise calculations on smaller molecular or on those systems with more complex chemistry, noted the SAB.

Because the results from many analytical techniques can be predicted computationally, simulation can be used to interpret and explain analytical results. This process can also be inverted so that a model of a particular structure can be developed from the analytical data.

"I believe that this synergy allows for an effective utilisation of both methods to determine and characterise molecular structure," said Tamara Zemlo, director of scientific and medical communications at the SAB, adding: "This combination approach will be instrumental in ensuring more cost-effective drug development in the pharmaceutical industry."