SafePathsRNAPC

This repository contains:

• The implementation of the computation of Safe Paths under different models of Path Cover in a Directed Acyclic Graph (DAG). C++ code.
• Experimental evaluation of the algorithm by computing contigs in a RNA Transcript Assembly problem. Jupyter Notebook.

A comprehensive explanation of both parts can be found in https://doi.org/10.1109/TCBB.2021.3131203.

C++ algorithm

First clone the repo:

 git clone https://github.com/algbio/SafePathsRNAPC.git

This project is a CMake project. To build this project with some runnables you should do

cd SafePathsRNAPC
mkdir build
cd build
cmake ..
cmake .. # Issue: second cmake necessary to compile external library
make

This C++ project downloads the LEMON graph library, which is stored in a Mercurial repository. As such, the installation requires Mercurial.

Jupyter Notebooks

After compiling the C++ code you can replicate our experiments by running the Jupyter Notebooks in the folder data. These notebooks create intermediate files in the different subfolders of data. The notebooks are self-contained and must be run in the following order (indicated in the notebooks too):

• data_manipulation/graph_creation.ipynb
• experiments/run_experiments.ipynb
• evaluation/compute_metrics.ipynb
• evaluation/compute_tables.ipynb

These notebooks correspond to the experiments for Homo sapiens. The experiments for other species can be found (following the same structure) in the folders data/mouse (Mus musculus), data/triticum_aestivum, data/hordeum_vulgare, data/fruit_fly (Drosophila melanogaster) and data/magnaporthe_oryzae.

Once all these notebooks have been run, you can run the notebook compute_summary_tables.ipynb.

The package requirements to run our jupyter notebooks can be found at data/requirements.txt

The datasets used are two BED files (per dataset) created from GFF files in the Enssembl project. The script used to transform the GFF into BED files can be found at data/scripts/gtf2bed.py

Contact

Any error, improvement or suggestion you can write me to elarielcl at Gmail.