This is the implementation of project Handbag. It finds and quantifies the directionality of intermolecular interactions. This project was developed for the thesis that I did for 't van het Hoff Institute for Molecular Sciences at the UvA.
Supervisor: dhr. dr. T.J. (Tiddo) Mooibroek Author: Natasja Wezel (BSc)
It takes as input orthogonal data from fragments. This data can be obtained from the Cambridge Structural Database, using ConQuest. You have to specify the central group and contact group, and the program will calculate the directionality coefficient. In the figure below, the steps of the program are given.
- Step 0: the data needs to be obtained by doing a conquest search. That data can be fed into the program; the labels need to be specified to be able to do the superimposition
- Step 1: Superimposition is done
- Step 2: The central group model is made by averaging each atom with the same label.
- Step 3: The densities are calculated at different bin sizes ranging from 0.2A to 1A.
- The density can be inspected in a plot with two sliders: one for threshold (what bins to show), and one for resolution (how big the bins are)
- Step 4: A reference point the contact group has to be specified (a single atom, or a centroid) and for all these reference points, the distance to the closest atom of the central group model is calculated. If this contact reference point does not meet the criterion distance
$\leq$ vdw distance + 0.5, the points are thrown away.- Fingerprint plots of the contact pair are made after this step
- Step 5: The volumes are calculated, and with this, the directionality score is calculated, which was the ultimate goal.
This GitHub project contains the source code for all separate steps in separate scripts, and also contains a Command Line Interface which makes it possible to walk through the pipeline step-by-step. For a detailed tutorial on how to use the CLI click here.
This is a GIF showing one of the 4D density plots made with this program:
For installation guidelines, click here.
The code for easily running the pipeline on lots of contact pairs is found in the jupyter notebooks.
The original data is found here # TODO: where to store the data so it can be downloaded?
Pull request templates are provided. Issues and pull requests are always welcome. For other questions, you can always contact me at natasjawezel@gmail.com!