Emilio Benfenati and Alessandra Roncaglioni from the Mario Negri Institute for Pharmacological Research in Milan, Italy, explain why environmental and pharmaceutical chemists should pop into the computer lab

Computers are reshaping the way we explore the life sciences. A palette of in silico tools is offering different advantages which could transform environmental and drug research.

What are the advantages of computer methods? Computer experiments are much faster than in vivo and in vitro ones. Some computer programs can process thousands of compounds simultaneously in a matter of hours, or even less. Computer programs are typically cheaper, in some cases free. But there are also expensive programs, which require skilled personnel. Computer programs do not use animals, which meets ethical concerns about experiments with animals. Computer methods may only require a computer and internet access, making them more widely available than any other method in principle. They do not use chemicals, do not produce chemical waste, and can be used without a sample of the compound to be tested. Indeed, they can be used before the substance has even been made, using only the structure.

What kind of information can computer methods extract? There are different tools, ranging from detailed docking studies, which look at the interactions between drugs and their receptors, to models that examine how structure affects a drug's activity. 

In the case of the oestrogen receptor, for example, it is now possible to assess how different compounds interact with the receptor, identifying which chemicals are more likely to bind. This binding is of interest to pharmaceutical companies, interested in new drugs acting through the receptor. Computer programs can also obtain other useful information, such as identifying the part of the molecule that is responsible for a biological effect or its modulation. 

An in silico experiment: a compound's oestrogen receptor binding properties can be modelled accurately with computers

Oestrogen receptor binding information can also be used in the field of environmental and human protection. The new European REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals) legislation requires that each chemical that is produced on the market in an amount exceeding one tonne per year be characterised. This includes chemicals that disrupt the endocrine system - the organs that release hormones, such as oestrogen, in our bodies. Indeed, a series of substances, present in the environment and in our food, can induce activities that disrupt the endocrine system. It would be very useful to identify these chemicals before their marketing and use. Computer methods are attractive for this, potentially providing fast and cheap chemical screening. REACH requirements should therefore boost research on computer models for endocrine disruption effects.

How reliable are the predictions? Computers allow us to predict which substances will have receptor binding properties. This has been shown to be useful for identifying new drugs, reducing the numbers of compounds that need to be screened. However, the prediction is never 100 per cent correct. In predicting potential environmental endocrine disruptors, the main problem is the lack of suitable examples that can be used to provide the correct information. Unfortunately, most of the studies have looked at substances' binding activities, which is only one component of the process that leads to the final effect in the organism. The lack of wider-reaching data makes it difficult to assess the complete process. However, this is a consequence of a broader inadequacy in experiments to assess the phenomenon in vivo, which calls for more research in the area.

How useful can computer programs be in the future? A more comprehensive understanding of the endocrine disruptor field requires new perspectives. All sources of information need to be integrated in a more structured process: in vivo, in vitro and in silico studies must be used together to take advantage of the positive features each provides.

Read Benfenati and Alessandra Roncaglioni's tutorial review 'In silico-aided prediction of biological properties of chemicals: oestrogen receptor-mediated effects' in issue 3, 2008, of Chemical Society Reviews.