reactIDR 2.0.3

reactIDR: evaluation of the statistical reproducibility of high-throughput structural analyses towards a robust RNA structure prediction

• Published in BMC Bioinformatics

reactIDR is a Python package that evaluates statistical reproducibility across replicated high-throughput RNA structure profiling data (e.g., PARS, SHAPE-Seq, icSHAPE, DMS-Seq) to robustly infer loop and stem probabilities.

📥 Input

• Read count data (tabular format)
  • PARS
  • SHAPE-Seq
  • icSHAPE
  • DMS-Seq (assumed to enrich A/C only)

📤 Output

• Posterior probability for each site:
  • Loop (signal enriched in "case")
  • Stem (signal enriched in "control")

🧠 Algorithm

• IDR (Irreproducible Discovery Rate)
• Hidden Markov Model

🔧 Requirements

python >= 3.9
numpy >= 2.0.2
scipy >= 1.13.1
pandas >= 2.2.3

Optional packages for visualization:

seaborn
jupyter notebook

🚀 Installation

pip install reactIDR

▶️ Getting Started

Test datasets are provided in the example and csv_example directories. To run a demo using CSV input:

git clone https://github.com/carushi/reactIDR
cd reactIDR/csv_example
python -c "import reactIDR; reactIDR.run_reactIDR([
  '-e 0',
  '--csv',
  '--global',
  '--case', 'case.csv',
  '--output', 'test.csv',
  '--param', 'default_parameters.txt'
])"

📚 More usage examples and options are available in the Wiki.

🛠️ Scripts

📖 Reference

• R. Kawaguchi, H. Kiryu, J. Iwakiri and J. Sese. "reactIDR: evaluation of the statistical reproducibility of high-throughput structural analyses towards a robust RNA structure prediction" BMC Bioinformatics 20 (Suppl 3) :130 (2019)" ー Selected for APBC '19 proceedings

• Docker image for reactIDR

  • Convert bam to read count data
  • Find scripts and how to use at https://github.com/carushi/RT_end_counter

• IDR

  • Li, Qunhua, et al. "Measuring reproducibility of high-throughput experiments", The annals of applied statistics, 2011.
  • IDR in Python
  • IDR in R

TODO

• apply to MaP analyses