Software

Bespoke software from the lab.

pyCRAC Software

pyCRAC is a comprehensive software package designed for the analysis of high-throughput sequencing data derived from CRAC (Cross-linking and Analysis of cDNA) and CLIP (Cross-linking and Immunoprecipitation) experiments. It provides a suite of tools for read preprocessing, mapping, and downstream analysis to identify RNA-protein interaction sites with high resolution.

Note that the latest version of pyCRAC is compatible with Python 3. Version 2.0 is on the way, which will feature massive improvements in performance and usability.

Installation

Install the latest stable release from PyPI:

pip install pyCRAC

Install from Bioconda:

conda install bioconda::pycrac

Install the latest development version from GitLab:

git clone https://git.ecdf.ed.ac.uk/sgrannem/pycrac.git
cd pycrac
pip install .

A software package without bugs does not exist. Please check the repository regularly for bug fixes and other updates. Feedback is also much appreciated!!

Please cite the following paper when using the package for your data analyses:
Shaun Webb, Ralph D. Hector, Grzegorz Kudla and Sander Granneman.
PAR-CLIP data indicate that Nrd1-Nab3-dependent transcription termination regulates expression of hundreds of protein coding genes in yeast.
Genome Biology 2014;15:R8; DOI: 10.1186/gb-2014-15-1-r8

GTF2 Parser Examples

Here are a few examples of how to use GTF2 parser in your python programs.

from pyCRAC.Parsers import GTF2
from pyCRAC.Methods import translate
import numpy as np
import re

gtf = GTF2.Parse_GTF()
gtf.read_GTF("Saccharomyces_cerevisiae.R64-1-1.75_1.2.gtf",transcripts=False)
gtf.read_FASTA("Saccharomyces_cerevisiae.R64-1-1.75.fa")

# Get genomic coordinates for gene RPL7A
gene = "RPL7A"
coordinates = gtf.chromosomeGeneCoordIterator("I")

See the full documentation in the package.

DBPeaks

DBPeaks is a Python package for identifying peaks in CRAC/CLIP data.

It uses pyCRAC tools together with bedtools and DESeq2 to identify reproducible protein-RNA interaction sites (peaks) in CLIP/CRAC data. It was specifically designed to compare CLIP/CRAC datasets generated under different conditions to identify differential binding sites.

DBPeaks was used in our recent Esteban-Serna nucleic acids research paper (Esteban-Serna et al. 2025) to identify differential binding sites (see Figure 1F).

DBPeaks Code repository

Reference

Sofia Esteban-Serna, Tove Widen, Mags Gwynne, Iseabail Farquhar, Michael Duchen, Peter S. Swain and Sander Granneman#.
A transcription termination mechanism for maintaining homogeneous protein expression.
Nucleic Acids Research. 2025 Nov 13;53(21):gkaf1199. DOI: 10.1093/nar/gkaf1199


 

Figure 1 from Esteban-Serna et al. 2025 showing DBPeaks usage

pyRBDome

pyRBDome is a computational pipeline developed to analyze RNA-binding proteome data. Over the past few years, we and others have produced a large number of proteomics datasets that revealed many putative RNA-binding proteins (RBPs). However, it was unclear what fraction of these putative RBPs were genuine RNA-binders.

To address this, we developed pyRBDome to enhance the experimental data. It is a comprehensive pipeline that enables users to identify putative RNA-binding regions within proteins and perform statistical analysis to determine if peptides identified in these proteomics studies are indeed bona-fide RNA-binding regions.

pyRBDome was also used in the following 2025 publications:

  • RNMT: L. A. Hepburn, Sander Granneman, V. H. Cowling and J. Clara Silva. RNMT is recruited to RNA by interaction with RNA G-quadraplexes. bioRxiv. 2025 Nov 16. DOI: 10.1101/2025.11.16.688685
  • NF45-NF90: Sophie Winterbourne, Uma Jayachandran, Juan Zou, Juri Rappsilber, Sander Granneman, Atlanta G. Cook. Integrative structural analysis of NF45-NF90 heterodimers reveals architectural rearrangements and oligomerization on binding dsRNA. Nucleic Acids Research. 2025 Mar 20;53(6):gkaf204. DOI: 10.1093/nar/gkaf204
pyRBDome Code repository
Figure showing pyRBDome predictions on Cas9 structure

Reference

Liang-Cui Chu*, Niki Christopoulou*, Hugh McCaughan*, Sophie Winterbourne, Davide Cazzola, Shichao Wang, Ulad Litvin, Salomé Brunon, Patrick J.B. Harker, Iain McNae and Sander Granneman. pyRBDome: A comprehensive computational platform for enhancing RNA-binding proteome data. Life Science Alliance. 2024; 7 (10) e202402787; DOI: 10.26508/lsa.202402787

Installation

We strongly recommend installing pyRBDome in a separate virtual environment to avoid conflicts with other packages.

Source (GitLab)

Create a virtual environment and install from GitLab:

# Create and activate a virtual environment
python3 -m venv pyrbdome_env
source pyrbdome_env/bin/activate

# Clone and install
git clone https://git.ecdf.ed.ac.uk/sgrannem/pyrbdome.git
cd pyrbdome
pip install .

diffBUM-HMM

diffBUM-HMM is a robust statistical modeling approach for detecting RNA flexibility changes in high-throughput structure probing data.

It provides a statistical framework for analyzing data from high-throughput RNA structure probing experiments (such as SHAPE-MaP, DMS-MaPseq, etc.) to identify significant changes in RNA structure flexibility between different conditions.

Reference:
Paolo Marangio*, Ka Ying Toby Law*, Guido Sanguinetti# and Sander Granneman#.
diffBUM-HMM: a robust statistical modeling approach for detecting RNA flexibility changes in high-throughput structure probing data.
Genome Biology, 2021. PMCID: PMC8157727 DOI: 10.1186/s13059-021-02379-y

# corresponding authors, * These authors contributed equally

diffBUM-HMM Code repository

Reference

Paolo Marangio*, Ka Ying Toby Law*, Guido Sanguinetti# and Sander Granneman#.
diffBUM-HMM: a robust statistical modeling approach for detecting RNA flexibility changes in high-throughput structure probing data.
Genome Biology, 2021. PMCID: PMC8157727 DOI: 10.1186/s13059-021-02379-y

# corresponding authors, * These authors contributed equally

MRSA Cellpose Biofilm analysis code

We developed a confocal image segmentation and fluorescence quantification pipeline for S. aureus biofilm Z-stacks. This Python pipeline detects and counts cells, measures pericellular ring fluorescence, and generates publication-quality figures. This pipeline was made to analyse CLSM data generated in the lab by Mehak Chauhan. 

The pipeline processes Nikon ND2 confocal Z-stacks across three independent datasets. For each sample it:

  1. Reads Z-stack and pixel-calibration metadata from the ND2 file
  2. Applies Richardson–Lucy deconvolution to sharpen the segmentation channel
  3. Detects in-focus Z-slices using a per-slice SNR metric
  4. Segments cells with Cellpose (cyto2 model, 2D+stitch). 2D+ stitch was used as this gave the best performance
  5. Filters objects by physical volume and shape
  6. Measures per-cell fluorescence intensity (cell body + pericellular ring)
  7. Saves per-cell CSVs, summary CSVs, deconvolved TIFFs, QC plots, and 3D renders
test
Three-dimensional visualisation of Z-DNA and B-DNA distribution in MRSA biofilms.
(a) Orthogonal (XZ) projection of a 72h biofilm stained with Hoechst 33342 (blue, total DNA), B-DNA (TOTO-3 iodide; magenta), and anti-Z-DNA antibody (red). The imagre was generated using Imaris.
(b) Focus quality metrics (signal-to-noise ratio) across the Z-stack for each channel. All three channels show asymmetric profiles with sharp signal onset and gradual decay, reflecting the optical sectioning characteristics of the biofilm. Peak positions are offset between channels, consistent with the vertical stratification observed in (a).
(c) Three-dimensional surface reconstruction of Cellpose segmentation results for a representative WT USA300 biofilm at 72h. The semi-transparent mesh was generated by marching cubes on the binary union of all segmented cell masks per channel. Colours indicate fluorescence channel: blue, Hoechst (nucleic acids); magenta, TOTO-3 (eDNA:B-DNA); red, anti-Z-DNA antibody (eDNA:Z-DNA). The spatial stratification along the Z-axis reflects the distinct focal depths at which each channel is optimally resolved, as determined by the per-channel SNR-based focus detection in (b). Axes indicate physical dimensions (µm). Note that cells appear elongated along the Z-axis due to the confocal axial point spread function (~1–2 µm) exceeding the physical cell diameter (~0.8 µm) at the 1 µm Z-step used; this does not affect cell counts or intensity measurements.
Biofilm Cellpose Pipeline link

Reference

Non-coding RNA RsaE regulates biofilm thickness, viability and dissemination in methicillin-resistant Staphylococcus aureus

Mehak Chauhan, Ivayla Ivanova, Emily G. Sudnick, Ryan W. Steere, Julia R. Tennant, Jacob A. Hensley, Pedro Arede, Gregory M. Jensen, Isabelle Hatin, Olivier Namy, Philippe Bouloc, Ronan K. Carroll#, Sander Granneman#
 
# Corresponding authors
 

BioRXiv 2026 March 6

This article was published on