A new analysis of drug screening libraries reveals a bias towards chemicals likely to be recognized by biological targets. The study, published online this week in Nature Chemical Biology, explains, in part, why these screening techniques have been successful in the past and why they are likely to have important implications for how to increase future success rates of drug discovery.
The dominant approach for finding new drugs, high-throughput screening -- HTS -- allows researchers to quickly conduct biochemical tests on millions of candidate drugs. The total number of possible chemicals that could be used in a screen is at least 1060. In contrast, a large HTS chemical library used in a pharmaceutical company might contain less than one billionth as many chemicals. With such a small sampling of the total 'chemical space' it would seem unlikely that HTS would work. But it does. Brian Shoichet and colleagues have developed a method to quantify the extent to which chemicals in screening libraries look like the chemicals found in nature and they find evidence that screening libraries are substantially biased towards these 'biological' chemicals.
The results suggest an important explanation for the current success rate of high-throughput screening and also provide clues as to why it fails for some targets. For these failures, the results suggest new chemicals, currently missing from screening libraries, could be added to increase the likelihood of finding drug leads.