10,000 Compounds and Counting: Scientists Create Extensive Chemical Library to Study Toxic Substances, Share Data

September 21, 2021

The Tox21 compound collection is managed inside an automated system operating at ⁻20°C. Samples are retrieved in individually barcoded test tubes. (Katlin Recabo, Glenn Gomba and Paul Shinn, NCATS)

We are exposed to tens of thousands of environmental and chemical agents, and testing their toxicity can be costly. The methods scientists employ typically evaluate only one chemical compound at a time over months or years. To speed up this process, NCATS scientists and their collaborators have created a large chemical library that they and others are using to probe the potential toxic effects of compounds quickly and efficiently.

The library is part of the Toxicology in the 21st Century (Tox21) program, a federal research collaboration among NCATS, the National Toxicology Program (NTP) at the National Institute of Environmental Health Sciences, the U.S. Environmental Protection Agency (EPA), and the U.S. Food and Drug Administration (FDA). The project aims to develop new ways to determine whether substances can adversely affect human health. The Tox21 10K chemical library combines libraries from NCATS, NTP and EPA into a set of approximately 10,000 compounds spanning environmental chemicals — known toxic substances — and FDA-approved drugs.

The Tox21 10K library enables researchers to perform rapid laboratory experiments evaluating how different compounds react within cells, alter chemical signals that are necessary for specific cell functions, or disrupt cell health. At NCATS, researchers do this through high-throughput screening. Because the Tox21 10K chemical library includes the NCATS Pharmaceutical Collection — a group of compounds representing more than 90% of all FDA-approved drugs and some drugs still in clinical trials — Tox21 investigators can test directly for toxic side effects of marketed drugs. Other compounds in the Tox21 library represent environmental chemicals to which people are exposed frequently and for which toxicity data are lacking.

“The Tox21 compound library represents the largest, most diverse chemical library ever constructed for the express purpose of probing chemical mechanisms of toxicity, with public release of all data a hallmark of the program,” said Ann Richard, Ph.D., chemistry lead for EPA’s Tox21 group. “The consolidation not only of the Tox21 partner libraries but of the agency-specific capabilities and varied programmatic interests effectively expanded the scope and impact of the screened library well beyond what each agency could have achieved separately.”

Ruili Huang, Ph.D., informatics and computational toxicology lead for NCATS’ Tox21 group, is eager for people to use Tox21 data to build models that can predict the toxicity of substances before they are tested in a laboratory. According to Huang, Tox21 investigators “generate a lot of data, and then [they] make those data public. Researchers who do computational modeling and want to make predictions [need data to do so], and this provides a very high-quality data set to train their computational models.”

To facilitate the development of predictive computational models, the Tox21 team held a competition, the Tox21 Data Challenge 2014, in which scientists worked to develop the most accurate predictive models using the Tox21 10K chemical library data. More than 100 teams worldwide participated, and the top-scoring models were 80% to 90% accurate in their toxicity predictions.

“The Tox21 project also has impacted federal regulatory processes,” said NCATS scientific director Anton Simeonov, Ph.D. For example, Tox21 researchers established tests assessing the effects of various compounds on the estrogen receptor signaling pathway — the cellular communication track that lets the body know how and when to respond to the hormone estrogen. These assays have been adopted by regulatory agencies and are included as part of the recommended testing array evaluating the effects of chemicals on the estrogen receptor.

In addition, the Tox21 project has provided lessons for constructing complex chemical libraries. The Tox21 team has established best practices and information that will be instrumental in future studies requiring the evaluation of highly diverse compounds.

“We want to streamline the chemical testing and still tell how it may be dangerous to people or say with some degree of confidence that it will be safe. We’re getting there, and generating this 10K data is leading us in that direction,” said Simeonov.

Learn more about Tox21 on the NCATS website and in an American Chemical Society review article.