STKE

Thinking Inside in the Box

+ See all authors and affiliations

Science  22 Aug 2003:
Vol. 301, Issue 5636, pp. 1021
DOI: 10.1126/science.301.5636.1021b

In attacking diseases whose molecular mechanisms are not known, devising an assay for a high-throughput screen in drug development is particularly challenging. Gunther et al. describe a screen that uses microarray analysis of gene expression to classify drugs that are currently used to treat depression or psychosis and those that interact with opioid receptors, which are important in the treatment of pain. They used two supervised classification methods—the classification tree (CT) and random forest (RF) algorithms—that were trained on the microarray data. Primary cultures of human neuronal cells were treated with antidepressants belonging to one of four subclasses [atypical, tricyclic, monoamine oxidase inhibitors, and serotonin selective reuptake inhibitors (SSRIs)]; classical or atypical antipsychotic agents; or opioid receptor agonists for the δ, κ, or μ opioid receptors. Thirty-two of the 36 drugs were correctly classified by CT and 30 by RF. Furthermore, RF analysis categorized SSRIs and tricyclic antidepressants into their subclasses when these were handled as unknowns. The marker genes identified as predictive for each drug class may provide insights into the molecular mechanism underlying the disease. Delta opioid receptor agonists were misclassified as antidepressants by both CT and RF, suggesting that this receptor might be a target for the treatment of depression. This approach may increase the efficiency of screening for new types of drug leads, especially for complex phenotypes where multiple pathways may converge to the disease state. — NG

Proc. Natl. Acad. Sci. U.S.A. 100, 9608 (2003).

Related Content

Navigate This Article