The back-end research focuses on extending retargetability by investigating reuse patterns and target descriptions.
* Properties, including the target description, that affect code generation.
3 : jhr 3864 * Properties, including the target description, that affect code generation.
Proteochemometric modeling [1] adds a target description, based on physicochemical properties of the binding site, to these models.
Given that both ligand- and target descriptors are used for PCM models, it follows that the target description is as important as the ligand description.
Given that ligand and target descriptors jointly form a PCM model, the target description is as important as the ligand description.
Limited literature is available using different approaches for target description but these are in most cases physicochemical properties similar to the z-scales, [5, 7, 19].
Several publications are available using varying ligand descriptors [7, 12, 13], yet on the side of target description there is less literature available.
If the target description included the word "enzyme", "cytosolic", "receptor", "agonist" or "ion channel" and the bioactivity record of the compound contained an IC50, Ki or Kd < 500 μM or had an activity > 50% binding affinity, then it was selected.
An overview of the 88 data sets with the target IDs, target description, number of actives, number of decoys, ratio actives/decoys, number of Bemis-Murcko scaffolds (BMS) [41] found in the actives and ratio BMS/actives is given in Table 2.
