Our genetic make-up dictates eye and hair colour, how quickly we age and what diseases we might acquire. Researchers are also now learning that genes influence the effectiveness of the medicines we take.

While pharmaceuticals save thousands of lives every day, not all people respond to them in the same way. An Irish research company, SurGen, based at the Royal College of Surgeons in Ireland is trying to understand the interaction between our genes and the drugs used to control illness.

"What we are interested in is the genetics of common disease," said Prof Desmond Fitzgerald, director and principal investigator at SurGen. "We are studying large populations from large drug trials, looking for the links between genes and disease."

The company was established in 1998, a fifty-fifty partnership between the Royal College and a European genomics company, Genset. It employs seven in a £1.5 million laboratory set up by SurGen. The college provides research and clinical expertise and Genset supplies technology and an existing knowledge data bank in an area known as pharmacogenomics, the study of how a person's genes influence the actions of a drug.

It's main clients are the world's leading pharmaceutical firms and most of the companies it works for are among the top 10 in this sector. Only a handful of companies can offer the type of service sold by SurGen, said Prof Fitzgerald.

Trying to link genes with drug activity is a difficult undertaking given the complex interaction between genes. Illness is usually caused by changes in large groups of genes not just one or two. Our genes are like small chemical factories. Each contains a recipe written in DNA for a particular chemical needed by our cells to live and grow.

Equally important to having the correct recipes is being able to regulate these factories. They should only produce their chemicals when they are needed and must switch off when enough has been produced. This is a finely balanced process and when it goes out of kilter then disease is the usual result.

SurGen is interested in genes that play a role in disease and the way that the substances produced by the genes interact with pharmaceutical compounds. "It is difficult to identify what genes are involved because there are 140,000 different genes and up to 50 variations for each one," said Prof Fitzgerald.

The variability factor is particularly important, he said. While we might all have a particular gene, there can be small differences from one person to the next and this variability can affect the way a drug will be utilised by the body. The variability is known as "single nucleotide polymorphisms", said Prof Fitzgerald. With such a vast number of genes available to study, the search for those linked to a particular disease must be focused down tightly. Biochemical studies can show what cell chemicals are in over or under supply in a disease state and this leads the researchers back to the genes which produce these substances. The SurGen team looks closely at the functional regions of the gene, and searching for the small variations between patients.

The key is to have large numbers of patients whose disease state and physical condition is well known and documented and this is done by linking up with international drug trials, said Prof Fitzgerald. The company has access to the data coming from up to 950 drug trial centres around the world, where a group of patients with a known disease state are given specific drugs.

Strict confidentiality applies and SurGen does not need to know the identity of the patients. But the company does get access to the detailed profiles done on each patient and to samples of genetic material. This represents a very powerful resource when trying to link disease status and genetics.

The work done at SurGen leads in several directions, said Prof Fitzgerald. It helps to identify what genetic combinations leave one more likely to become ill and by extension. "If you could find the at-risk population you could target a pharmaceutical therapy," he added.

These studies also provide information about the genetic component of a patient's response to a drug. Most drugs work for most people but there are a small number who are completely unresponsive to a therapy. There are also classes of drugs, for example those used to reduce the risk of blood clots, which only work in about one in four patients.

This kind of information can feed directly back into clinical practice so that improvements can be made in drug therapy. "We are helping to tailor who they give a drug to," said Prof Fitzgerald. The ultimate goal, he said, would be to design a drug therapy specifically suited to the genetic make up of a patient. "You might be able to develop genetics to the point where you could check a patient's genetics before they take a drug."