A linear model of water/octanol partition coefficients
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
model
phys
props
.gitignore
README.md
overlap.dat
testing.xlsx
training.xlsx
training_filtered.xlsx

README.md

On the use of linear models to predict partition coefficients

Simple and extensive linear models to predict water/octanol partition coefficients for small molecules.

Models

There are two models in two subfolders:

  • model - the extensive modelling efforts in the MSG500 course (autumn 2017)

  • phys - simple model presented at my docent lecture and at the Combine Engineering Journal

Experimental data

The experimental training data was collected from the SI from the following article: Automated Parametrization of the Coarse-Grained Martini Force Field for Small Organic Molecules. The Table S1 was manually copied to the file training.xlsx

The experimental test data was collected from the Minnesota database and copied to the file testing.xlsx. This is a subset used in the paper "Solvation free energies and partition coefficients with the coarse-grained and hybrid all-atom/coarse-grained MARTINI models".

Collecting molecular properties

The properties of the molecules were collected using the Indigo package

python props/collect_props.py training.xlsx

Overlap of experimental data sets

The indices of the molecules in the training data that is also in the test data was created by

python props/compare_smi.py training.xlsx testing.xlsx

and the indices are available in overlap.dat

The experimental training data without this overlap is available in training_filtered.xlsx.