All the equations based on descriptors of different categories involved quantum-chemical descriptors (Table S6).
All equations based on descriptors of the same categories used quantum-chemical descriptors, and in particular EHOMO and ELUMO.
The equations are mostly based on one descriptor, but they can involve up to 27 descriptors.
The 40 equations allowing the estimate of vapor pressure involve from 1 to 12 descriptors, and more than half of them are only based on one descriptor.
First, it considers all training set compounds, selects the descriptor that is the best predictor of the value of the target variable (a tissue's log Kt:p in the first phase or log Vss in the second phase, in this work), adds the selected descriptor to the tree (as its root node), and partitions the training set based on the values of the selected descriptor.
While the Z-scale descriptors introduced originally by Wold [1] are based on experimental values, we have applied computational amino acid descriptors [2] derived using the CoMFA method [3] for building QSAR models, which are amenable to general amino acids.
Whereas the curves from models with 2D descriptors differed across the entire range of X values, the models based on 3D descriptors showed large differences at X values between 0 and 0.4, then reached S X) = 1 at larger values of X.
To model heterogeneous residual variance across the environments, the observations were ordered according to the environmental descriptor and grouped into 10 classes of equal size based on the environmental values.
