Code for "Genome-wide modelling of transcription kinetics reveals patterns of RNA processing delays"

This repository contains code used to produce the results in the paper "Genome-wide modelling of transcription kinetics reveals patterns of RNA processing delays".

All code is available under the BSD license (see LICENSE.txt).

Gaussian process (GP) modelling code

The GP code is implemented in Matlab and uses the GPmat library (https://github.com/SheffieldML/GPmat).

The main script for running the code is 'matlab/run_all_gpnddisim_hmc_convcheck_mixprior.m' for the real data and 'matlab/run_all_gpnddisim_hmc_convcheck_mixprior_synthetic.m' for the synthetic data.

The scripts make assumptions about file names and locations so they would need editing before running.

HMC sample post-analysis code

There is a number of Matlab scripts for post-processing the HMC samples: 'matlab/check_early_profiles.m' 'matlab/combine_profiles.m' 'matlab/analyse_hmc_new.m' 'matlab/get_hmc_medians_new.m'

The scripts make assumptions about file names and locations so they would need editing before running.

Model plotting code (cf. Fig. 1)

The main entry point is 'matlab/run_plot_sampled_predictions.m'.

Analysis of genomic features associated with delays

The analysis code is written in R. The main script is 'R/analyse_intron_lengths_meddelay.R'. The script run the analysis and writes most of the figures in the paper.

pre-mRNA distribution analysis

The code is written in Python and is at 'python/plotwigs.py'.

Data pre-processing

Pol-II pre-processing code is in 'matlab/pol2_processing'. RNA-seq pre-processing code is in 'matlab/rnaseq_processing'.