Introduction

This repository contains the code to reproduce the experimental results in the paper "Base-pair Ambiguity and the Kinetics of RNA Folding".

Setting up the environment

1. Clone this repository

git clone https://github.com/StannisZhou/rna_statistics.git

2. Set up the virtual environment

cd rna_statistics
conda env create -f environment.yml
source activate rna_statistics

Reproduce the results

Use

streamlit run main.py

to reproduce all the results using cached results in the intermediate folder. This would open up a browser session and interactively generate all the supplementary results.

A pre-generated report containing all the results is also available at report/index.html.

Regenerate all the results from scratch

To facilitate the exploration of the results, we included cached results in the intermediate folder. To regenerate all the results from scratch, a few additional external dependicies are needed:

1. Use the standalone implementation of Knotted to Nested to remove pseudoknots from the comparative secondary structures available under data. A copy of the pseudoknot-free comparative secondary structures (generated using the IR option) is available under data_without_pseudoknots. Note that you need to make simple modifications to the bpseq parser in the standalone implementation of Knotted to Nested so that it works with the data available under data, and use the code under python 2.7.
2. Set up the RNA Secondary Structure Analyser, available at the RNA STRAND database website. You might need to change line 42 of interval.cpp, and properly set X11 related options in Makefile (e.g. properly set X11_INCLUDE and change X11 related options in LFLAGS) in order to successfully compile the analyser. The code assumes the analyser is available in the RNAAnalyser folder under you home directory. If it's placed in a different place, please change line 14 of data.py accordingly.
3. Install the Vienna RNA package. Make sure RNAfold and RNAPKplex are available at the commandline.

After setting up these external dependicies and generate the data_without_pseudoknots folder, you can remove the intermediate folder and use

python run_experiments.py

to regenerate the cached results in intermediate.