As accessible compound catalogs continue to expand, the question of how to reliably extract relevant chemistry from billions to trillions and more of entries is becoming increasingly important. In particular, the often intransparent understanding of molecular similarity and degrees of relatedness makes it difficult to compare ranking lists or similarity algorithms with one another. In this webinar, we explore the extraction of relevant molecules from ultra-large Chemical Spaces. Building on our recently published study, we discuss both the opportunities and challenges of computational ligand-based screening, as well as the medicinal chemistry rationale behind different similarity metrics.
