Researchers at MIT and Harvard have released a public database that they hope will be the seed of a new human genome project, one that lets researchers find potential treatments for disease in a specialized search engine.
Both diseases and drugs have so-called genetic signatures—sets of genes known to be turned on or off, up or down in patients with a particular disease or in cells exposed to a particular drug.
The database project, termed Connectivity Map, lets researchers correlate signatures from diseases and drugs, essentially suggesting drugs that may treat diseases for which they had not previously been considered.
“The Connectivity Map works much like a Google search to discover connections among drugs and diseases,” said senior author Todd Golub, the director of the Broad Institutes Cancer program.
Drugs originally developed to treat one disease are frequently found effective against other diseases through chance and accident.
Common antidepressants, for example, started out as antibiotics and antihistamines. The impotence drug Viagra was originally studied to treat cardiovascular disease.
The Connectivity Map can help researchers use computers to find such unanticipated possibilities. Thats much faster than waiting for clinicians to notice odd side effects.
“By analyzing just a small fraction of available drugs, we have already confirmed several biological connections between drugs and human disease, and made entirely new ones, too,” said Justin Lamb, a senior scientist in the Broad Institutes Cancer program.
Nearly any researcher with an Internet connection can access the Connectivity Map. Genetic signatures associated with a disease, drug, or other condition are entered as search terms. Compounds in the database with corresponding signatures are then displayed in order of how close the match is.
One previously unknown connection is between a chemical extract of a type of cedar tree and cancer. The extract, called gedunin, had been studied to treat malaria and other diseases.
The database found gedunin in a screen for molecules that disrupted hormone signals in prostate cancer cells, then used that information to find out precisely how it affected cells.
So far, the project holds signatures for about 160 drugs and “bioactive” molecules, but researchers at the Broad Institute, a joint venture between Harvard and MIT, want to expand it.
Data from the study are made publicly available here. A Web-based tool for scientists to perform their own analyses using the Connectivity Map is also freely available at this site.
Results of studies using Connectivity Map were published in Science and Cancer Cell on September 28 and 29.