A new path has opened for finding drug candidates without testing countless compounds one by one. Korean researchers have launched a public drug discovery service that uses DNA sequences as "barcodes" to simultaneously screen and validate tens of millions of compounds. Until now, domestic researchers and biotech firms often outsourced large-scale compound screening to overseas companies at high costs, but they can now access such services at home.
The Korea Research Institute of Chemical Technology (KRICT) said its DEL Technology Research Group has launched the "DEL Core Bank Platform," which supports drug discovery based on DNA-Encoded Library (DEL) technology. DEL is a technology that attaches a unique DNA sequence to each compound, making it identifiable. The DNA serves as both a name tag and a barcode for each compound.
Drug development begins with finding substances that bind well to or react with disease-related proteins. Such substances are called active substances or hit compounds. For example, if there is a protein involved in cancer cell growth or a protein that triggers inflammatory responses, researchers must find substances that can bind to these proteins and block or regulate their actions. This process is the first hurdle in drug development.
Previously, high-throughput screening (HTS) was mainly used to find active substances. HTS divides compounds one by one into multiple small experimental wells and checks whether they react with the target protein. Because reactions can be observed directly without additional treatment of the compounds, the results are highly reliable. However, as the number of compounds to test grows, so do the time and cost burdens. For example, physically inspecting 60 daily plates of 384 wells containing 1 million compounds takes 43 days. Including preparation and analysis, the process takes about two months.
The DEL technology unveiled by KRICT is useful because it allows numerous compounds to be tested together rather than individually. The device that makes this possible is the "DNA barcode." When researchers create compounds, they record in the DNA which chemical fragments were used. As a result, even when numerous compounds are mixed together, reading the DNA information reveals the structure of each compound. For example, 100 types of chemical fragments are prepared, and each fragment is given a different DNA barcode. These fragments are then combined with other chemical fragments, and that information is also recorded in the DNA. Repeating this process exponentially increases the number of compounds with different structures. Combining 100 types of chemical fragments can produce 1 million different compounds.
The numerous compounds created in this way meet the target protein while mixed together in a single solution. The target protein is a disease-related protein that researchers seek to regulate with a new drug. Compounds that bind well to the target protein remain as candidates, while those that do not are washed away. Researchers then read the DNA barcodes attached to the remaining compounds to identify which compounds bound well to the target protein. With this method, screening can be completed within 15 days to one month even when the number of compounds exceeds tens of millions.
However, when observing reactions in a solution containing many mixed compounds, errors can occur in which substances that are not actual candidates happen to be filtered together, or specific DNA sequences are read disproportionately often. To reduce such errors, KRICT uses artificial intelligence (AI) analysis. The approach involves training AI on extensive experimental data to identify patterns of which chemical structures bind well to target proteins, and then re-selecting candidates with high potential.
KRICT plans to select the top 50 substances most likely to develop into drug candidates from those screened and provide a final report. Rather than simply providing a list of compounds, the institute will establish a one-stop service system covering library provision, screening, data analysis, follow-up synthesis, and validation.
KRICT built the platform with support from the Ministry of Science and ICT and the National Research Foundation of Korea under the "Core Bank Construction Project Based on a Large-Scale DNA-Encoded Library Platform." Accordingly, companies and institutions using the service will receive a temporary 50% discount on costs through 2027. Daewoong Pharmaceutical, iLAB, the National Cancer Center, Ewha Womans University, and the Gwangju Institute of Science and Technology (GIST) are currently receiving priority support. The service is expected to function as public infrastructure that reduces the time and cost burden of large-scale compound screening, broadening the foundation of Korea's drug development ecosystem. "We will lower the foreign dependency of DEL technology and provide support so that everything from initial active substance discovery to follow-up validation can be carried out efficiently within Korea," said Heo Jung-nyoung, head of KRICT's DEL Technology Research Group.
