Introduction to Docking

Typically, the docking of a molecule generates several possible orientations, the so-called ‘poses’. On the top several possible binding modes of the painkiller paracetamol/acetaminophen are presented. Several binding modes are usually offered after a docking process to the user for visual assessment to avoid accidently missing out a good prediction. Good means here that the molecule forms favorable interactions with the target thus increasing the stability of the complex. This is important because stable complexes are more likely to occur and therefore are also to be expected in reality.
But it leaves the person with a question: Which pose is the correct one?
This is where scoring is applied.

Scoring assesses the interaction quality of the complex (ligand and target) of a pose. Poses that display strong interactions with residues of the target are more likely to be observed in reality. Therefore, scoring is used to decide which pose is the best one as it can be viewed as a surrogate parameter for the binding affinity of the compound. Given the fact that scoring provides a numerical value to each pose, it becomes possible to rank poses.
Ranking allows comparison of individual poses of a single ligand, or the comparison of binding modes of different ligands. The ligand with the better scoring is more likely to have a better affinity and therefore more likely to bind which is the premise for any pharmacological effect at the target. The latter is also the premise of virtual screening: A method were a virtual set of molecules are docked at a target and scored for their binding affinity. It can be used to computationally narrow down which compounds are likely to bind at the target structure which saves time and resources compared to handling thousand or millions of physical compounds. The top scoring compounds are assessed for their properties candidates and subsequently selected for acquisition and pharmacological evaluation in assays.
And although the scoring functions have improved in the past decades they are still far from being perfect. The holy grail of a scoring algoritihm that can reliably predict the activity of a compound with a 100 % success rate is yet to be found. Consequently, medicinal and computational chemists still rely on their experience and supporting information (e.g. molecular torsions, intra- and intermolecular clashes) to assess binding mode predictions.

FlexX is the docking algorithm of SeeSAR. It places a part of the molecule in a defined binding site and builds up the whole structure based on binding predictions of the initial fragment. This speeds up the docking progress in comparison to docking the whole molecule at once allowing to investigate more poses with less effort.

HYDE is the scoring algorithm of SeeSAR. It assesses the binding affinity of any ligand in a binding site, be it from docking or from a crystal structure. SeeSAR is also able to visualize HYDE in the graphical user interface: Atoms that contribute well to the binding affinity receive a green sphere, atoms that lower the ability of the compound to bind due to non-existing or bad interactions receive a red sphere. The size of the sphere reflects the quantitive contribution to the overall binding affinity of the ligand. By loading a structure to SeeSAR users have therefore the ability to visually assess where improvements can be made to make the compound more potent.