With the size of commercial libraries now reaching several billions of compounds, screening vast chemical spaces is becoming an attainable prospect. Yet, current screening strategies are unsuitable for such large compound collections. Hence, new technologies need developing to maximise this chemical space and in turn improve our chances of identifying high-quality hits during virtual screening campaigns.
Our solution is a newly developed ultra high throughput virtual screening workflow harnessing the power of Machine Learning with advanced cloud computing and cheminformatics. While virtual screening campaigns are traditionally carried out using low-information 2D fingerprints, our technology combines high-quality structure modelling with cutting-edge Active Learning. This unique platform enables us to identify, in a timely and cost effective fashion, high quality virtual hits from vast compound collections.
We present below how we have incubated and industrialized this technology within Concept Life Sciences by performing a virtual screening of BRPF1 inhibitors.