Fig. 2: Chemical space comparison for DEL training dataset, validation set (known binders to CK1α/δ), and blind assessment set screened for hit discovery. | npj Drug Discovery

Fig. 2: Chemical space comparison for DEL training dataset, validation set (known binders to CK1α/δ), and blind assessment set screened for hit discovery.

From: Evaluation of DNA encoded library and machine learning model combinations for hit discovery

Fig. 2

The output of t-distributed stochastic neighbor embedding (t-SNE) analysis performed separately for three DELs, MilliporeSigma (MS10M) DEL, HitGen OpenDEL (HG1B), and DOS-DEL (DD11M) are shown in (a), (b), and (c), respectively. The Broad CC is the blind assessment set of 140 K compounds used to predict hits by the ML models. The known binders or validation set include literature-curated hits and in-house set of binders to CK1α and CK1δ.

Back to article page