Discovery of Hit Compounds through Virtual Screening

The search for hit compounds is a crucial step in drug discovery research. Hit compounds are initial molecules that exhibit promising activity against a specific biological target. The quality of these hit compounds significantly impacts the efficiency and outcomes of subsequent compound optimization processes. Conventional high-throughput screening (HTS) involves the experimental evaluation of hundreds of thousands to millions of compounds, which requires substantial time and costs. Therefore, more streamlined search methods are essential.

   

Expanding the Virtual Compound Library to an Ultra-Large Scale

In recent years, virtual compound libraries have dramatically expanded. A virtual compound library is a database of compounds generated by a computer. Virtual screening technology enables the rapid evaluation of these virtual compounds on a computer before they are synthesized and experimentally tested. Eisai has generated virtual compounds that can be relatively easily synthesized from combinations of our synthetic intermediates and reagents, constructing a virtual compound library comprising billions to tens of billions of compounds. This approach has significantly expanded our unique chemical space (structural diversity of compounds).

Technological Advancements in Virtual Screening for Handling Ultra-Large-Scale Libraries

Virtual screening technology is rapidly evolving, particularly with advancements in docking simulations. Docking simulation is a method for calculating the strength of interactions between compounds and target proteins using their three-dimensional structures on a computer. This technology allows us to select promising hit compound candidates from virtual compound libraries. By utilizing active learning and performing large-scale parallel computations through cloud computing, we have established an environment that can efficiently analyze vast amounts of data and complete docking simulations of tens of billions of compounds within a realistic timeframe.

What Does This Technology Enable?

The establishment of ultra-large-scale virtual screening methods has dramatically improved the speed of screening. This advancement enables us to suggest hit compound candidates from virtual compound libraries consisting of tens of billions of compounds, which was previously impossible with conventional methods. Consequently, the likelihood of discovering new compound structures and highly active hit compounds has increased, potentially shortening the timeframe for compound optimization.

Utilization at Eisai

Eisai is advancing the utilization of virtual screening. For instance, we are identifying new hit compound candidates for specific cancer cells, and experimental evaluations are underway. In the field of global health, Eisai is also utilizing this approach to develop treatments for Chagas disease, one of the neglected tropical diseases. Additionally, Eisai has successfully and efficiently explored selective inhibitors for specific phosphodiesterases (PDEs), which are therapeutic targets, by combining machine learning models trained on internal and external data with docking simulations. With this approach, we are working towards the rapid development of treatments in collaboration with our partner, the National University of La Plata (Argentina).