After writing python code to calculate physical properties of polymer molecules in 2021, we are interested in cleaning up the code, addressing some calculation issues, and putting the code on GitHub. The code is written using an open-source cheminformatics package called RDKit. Prior to using RDKit, we had been using commercial software (Materials Studio, Chem3D) from 2014 to 2019.
The physical property of interest relates to hydrophobicity or the oil-like characteristics of polymers. Our method is inspired by the medicinal chemistry approach to describe drug-like molecules using partition coefficients. These coefficients, which are often referred to as LogP values, can be positive or negative. Positive LogP values indicate oil solubility while negative LogP values suggest water soluble molecules. Since the 1980s, the pharmaceutical industry has spawned many computational methods to calculate LogP.
Our method constructs SMILES strings for a short segment of a polymer. These SMILES strings represent 3D chemical structures using ACSII symbols. Then, we use RDKit to convert the SMILES string to a 3D molecule, optimize the conformation, and calculate the surface area (SA). Afterwards, we calculate LogP. The resulting ratio of LogP/SA has provided predictive capability in a number of collaborative projects. Since 2015, we have published 18 journal articles that use LogP and LogP/SA values.
Project Information Subsection
The goals of the project include the following:
1) Clean up code
2) Optimize calculation method
a. Reduce time needed for jobs
b. Automatically adjust to available hardware resources
c. Determine how many conformations are needed
3) Data output
a. Output graph or list values (csv etc)
b. Provide options for selecting axes for graph (x-axis could be number of monomer units (N), 1/N, logN, N^1/3, N^1/2 etc.)
4) Put code on GitHub
{Empty}
- Experience with Python
- Experience with or interested in learning Git and using GitHub
Milestone Title: Clean up the code Milestone Description: Currently, the code needs to be cleaned up and possibly reorganized or restructured. Completion Date Goal: 2022-06-30
Milestone Title: Optimize calculation method Milestone Description: While running the code, we have noticed that calculating LogP is relatively fast. However, calculating surface area (SA) is much slower with most of the computational time spent finding the best conformation (i.e. lowest energy). If possible, we would like to reduce calculation time by automatically adjusting to available hardware resources and finding a better way to conduct the conformation searching. Completion Date Goal: 2022-08-11
Milestone Title: Data output Milestone Description: Currently, the code outputs a graph. We would like options to save a graph, list values (csv etc), or select axes for graphing the data. Completion Date Goal: 2022-09-30
Milestone Title: Post code on Github Milestone Description: Get code ready and make available on Github. Completion Date Goal: 2022-10-30
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
Final Report
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
From the Project PI:
"So far, we have created The Hydrophobicity Project on GitHub to build a community of users. Sanders code is available on this site.
https://github.com/TheHydrophobicityProject
We have been testing and using the code since last October. A manuscript is in preparation."
From the student facilitator:
"This was the first project I worked on where I used version control. Now I can't live without it. This was also the first project in which I was creating a "product" for general consumption, rather than internal tools. It was useful to collaborate with people that weren't active coders on the project so I could get usability feedback off of which I could iterate. Those are two things I will keep with me regardless of the other technologies I use in my future projects."