10. Analysing Results 10.1 Analysing Results in DataWarrior

10.Analysing Results

10.1 Analysing Results in DataWarrior

After you have extracted your Alldata file, you can open up your results in DataWarrior.

There are several key parts to analysing your results, the main 4 are highlighted below. When performing the docking experiments, you inputted a sdf file containing several compounds, for each compound many poses are docked based on the conformations of each compound that were generated. You will be able to view the docked poses in your target protein in PyMOL (see section 10.2). Each row is a pose, and each pose has two values associated with it, “minimisedAffinity” and “RFScore”, these are generated by SMINA during docking.

The first red box (1.) indicates the “minimisedAffinity” column, this is the binding affinity results SMINA has calculated from docking for each pose. The second box (2.) is the RFScore column, this is the docking score associated with each pose.

Next in the filter area you will notice two sliders, one for the binding affinity and one for the docking score. You will want to assess these ranges: generally larger ranges indicate results may need filtering, whilst narrow ranges may mean that this value cannot really be used to discriminate between results.

Docking scores and binding affinities should always be taken with a pinch of salt, a binding affinity of -10 vs a binding affinity of -5 does not necessarily mean that the former pose is twice as good. Likewise, you should not discount all poses with lower docking scores or poorer binding affinities. You should use these values from docking as a way to support your analysis of your poses in PyMOL and to help back up any conclusions you have made about the success of each compound.

You can copy data from DataWarrior into Excel if you wish to analyse/examine some of the results in Excel and possibly make some tables of your top X poses or compounds.

When you view your results in DataWarrior you will notice that in the molecule name column your poses are simply named after the compound they were generated from, you may also notice that the compounds poses are not in order, i.e. the compound 2 poses may be listed before the compound 1 poses (as seen in this example). For us this is not much of an issue because there were only 2 compounds in the original sdf for docking, however in longer lists of compounds it can be harder to keep track of where one compound ends and another begins when analysing results (particularly in PyMOL, see section 10.2).

You will also notice that each pose has not been given an individual name in the file, this can be remedied by saving the sdf file in DataWarrior and naming the “compounds” (poses) by row number (as shown in section 6.4), this will give each pose a name (compound 1, compound 2 etc for pose 1, pose 2 etc). Before you save you may want to reorder your results (see section 6) in the same way you ordered the compounds when saving your sdf file for docking. For example, if you ordered your compounds by LogP value, then you can calculate this property in your Alldata sdf when in DataWarrior, and then order your docking results by LogP. This should fix the issue of your docking results being displayed out of order (unless some compounds had identical values for the property you ordered the list by).

When you open your reordered and newly saved Alldata sdf file (here we renamed the file when saving as “AlldataReordered”) you will now see a new molecule name column which names each pose, each pose will also still have the original molecule name column telling you which original compound it was generated from (see screenshot below).

In the red box highlighted above you can see where the compound 1 poses end and the compound 2 poses start. Row 36 (pose 36, named compound 36) shows it’s a compound1.sdf pose in the “molecule name 2” column.

You may want to make a note of where each compound’s poses end, so that you can easily keep track of this when viewing your results in PyMOL.

Also note in the status area you can see how many poses were generated in total (54 here in this example).