Two ML models for the Solubility Parameter: Gaussian Process
Regression (GPR, default) and Random Forest (RF), trained on the current MD database including the drugs, together
with a GPR density model (see ML Models)
FH Parameter, Miscibility and
Phase Diagram use the MD values of the database whenever available.
Otherwise, δ is predicted with GPR, and the unit volume is calculated from the PubChem density or the GPR density
model
Database
Add PLGA 50:50 (DOP 184) and P(Phe-alt-G) DOP 55, and further runs of P(Val-alt-G) DOP 55 and
P(Phe-alt-G) DOP 90. PLGA is renamed to PLGA 20:80
Correct the SMILES of BRP187, BRP201 and MF15
Correct the simulation protocol data of the drug–polymer runs
Remove duplicated alcohol runs and a mislabelled butanoic acid entry; relabel the DOPs of polyvinyl butyral
Regenerate the aliphatic acid data (decimal separator errors) and add six acids back to the compound list
Remove 3-methyl-2-pentanone (MD runs to be repeated) and three polymers without MD runs from the compound list:
Poly{cyclopentanediacetic acid-alt-[bis(2-oxazoline)]}, Poly{cyclohexanediacetic acid-alt-[bis(2-oxazoline)]} and
polytrimethylene terephthalate
FH Table: add the missing 1/T2 terms to the χ expressions of six entries
Improvements
FH Parameter: mole fractions can be entered as fractions such as 5/6
FH Parameter, Miscibility and Phase Diagram: a note in the input form explains where δ and the unit volume
come from
Clearer axis, legend and color bar labels in Miscibility and Phase Diagram
Bug fixes
Unit volume for DOPs without MD runs is now based on the volume per repeating unit of the stored runs
Fixed the averaged view, the AKBA lookup, the phase-diagram grid, the DOP check for polymers, mole fractions of
0 or 1 in the FH Parameter, the matching of SMILES and polymers to database entries, and the search for
Polypeptoid Gamma