Hierarchical Statistical Mechanical model

Biophysical prediction of protein-peptide interactions and signaling networks using machine learning.


This website presents the predictions and models described in:

Cunningham, Koytiger, Sorger, AlQuraishi. "Biophysical prediction of protein-peptide interactions and signaling networks using machine learning." Nature Methods (2020). doi:10.1038/s41592-019-0687-1


This website allows access to two parts of the paper. Primarily, this website is built to provide a network view of the associated interactions. In this view, we highlight the neighborhoods of inferred proteins and thier components using pie charts. In addition, we built in a protein structure viewer to facilitate interaction with the inferred energies associated with different structures.

Data / Code

Data and code are available in a number of ways. Code is available via Github for the model. A separate repository stores the code associated with the figures from the paper. Complete data are available via figshare (doi:10.6084/m9.figshare.10084745).


The model, analysis, and website were developed / implemented in the AlQuraishi and Sorger Laboratories within the Laboratory of Systems Pharmacology at Harvard Medical School. Questions, comments, concerns should be submitted via the GitHub repositories: github.com/aqlaboratory/hsm (model / data / code associated questions) or github.com/aqlaboratory/hsm-web (website associated questions).

This website is possible due to several external codebases. The structure viewer depends on the excellent 3Dmol.js. The network utilizes D3.js. Styling of the website primarily relies on the Bootstrap stack.

Funding: U54-CA225088 (NIH), P50-GM107618 (NIH) and W911NF-14-1-0397 (DARPA / DOD)

© 2020, Laboratory of Systems Pharmacology. All rights reserved.