🔐 cuf-cryptography - Chemical unclonable functions for decentralized cryptography

• Overview
• Software Requirements
• Installation Guide
• Data Sources
• License

Overview

This repository contains the data analysis pipeline, in the form of Jupyter Notebooks and Python files, for the following publication:

Anne M. Lüscher, Andreas L. Gimpel, Wendelin J. Stark, Reinhard Heckel, Robert N. Grass. Chemical unclonable functions for decentralized cryptography. Manuscript accepted.

Specifically:

• the script optimization.py runs a parameter grid search to determine optimal parameters for the Jaccard similarity
• the scripts downsampling.py and downsampling_constrained.py perform downsampling to identify the number of reads required for successful discrimination of experiments
• the Python files fileio.py, kmers.py, analysis.py, and filtering.py supply methods utilized by the aforementioned scripts
• the Jupyter notebook hashing.ipynb performs analysis using the MinHash and Fuzzy Extractor
• all other Jupyter notebooks are for data curation and visualization

Software requirements

The data analysis has been tested and performed on Windows 10 using Python 3.9.7. The following Python packages were used and are required:

numpy
pandas
scipy
biopython
pillow
reedsolo
datasketch
plotly

Installation guide

To clone this repository from Github, use

git clone https://github.com/fml-ethz/cuf-cryptography
cd cuf-cryptography
pip install -r requirements.txt

These steps should only take a few seconds to a minute, and no further installation is required. No special hardware or resources are required for use.

Data sources

Most intermediary files are supplied with this repository, in the form of .csv files. Due to their size, the sequencing data is not supplied with this repository, but is instead deposited on figshare. To generate filtered sequencing data using BBMap (v38.99), the Python script in the data directory may be used.

License

This project is licensed under the GPLv3 license, see here.