Hello, I'm writing a thesis about models predicted by I-TASSER and the corresponding energy caculated by Quantum Chemistry method. I have several questions about the Number of decoys and cluster density.
According to SPICKER, the rank of number of decoys and cluster density should be the same for the five models, as the REMC method indicated.
However, the result is as follows:
Name C-score Exp.TM-Score Exp.RMSD No.of decoys Cluster density
Model1: -2.42 0.43+-0.14 7.4+-4.2 5028 0.1417
Model2: -2.36 3419 0.1509
Model3: -3.07 2371 0.0739
Model4: -4.08 606 0.0271
Model5: -2.91 1772 0.0875
If ranked by number of decoys, then model5 should rank the fourth and model1 as the first. However, while ranked by cluster density, model 2 should rank the first. I really got confused by the ranking. And is there a formula to calculate the cluster density? I didn't find any in the paper about SPICKER.
I took the structure from the PDB database and predicted the structure by removing the homologues. So I also compare the TM-score, RMSD as well as the potential energy relative to the native structure.
The result showed that model2 is the closest to the native structure. In this case, it should be ranked as model1 instead of 2.
Looking forward to your reply & plenty of thanks.