We have previously described canSAR’s unique suite of machine learning algorithms for the prediction of ligandability of proteins (1). canSAR provides a visual summary of ligandability that is weekly updated. It presents approved drugs, clinical candidates and chemical probes that are curated in canSAR (Figure 2A). A domain representation of the protein sequence is displayed (Figure 2B) with a histogram (in green) showing the concentration of ligandable cavities along the sequence. The example of EGFR in Figure 2B shows that ligandability is clustered around the protein kinase catalytic domain. Figure 2C shows the assessment using canSAR’s orthogonal target-likeness assessments using precedence, protein network behaviour, and accessibility to antibodies.
