for some docking cases, the top cluster size is rather small (10) compared to the second cluster (128). The difference of z-score (-1.2 and -1.1) is not significant, however the RMSD deviated a lot (0.6 and 12.7). In such a case, would it be fine if we proceed with the top cluster? or we should consider the docked pose from the second class?
The large RMSD differences indicate the two clusters represent different solution. I would carefully examine both…