Intron Retention Alternative Neoantigen Prediction (IRANP)

IRANP is a pipeline designed to identify intron-retention–derived neoantigens by integrating RNA-seq analysis, HLA binding prediction, and peptide validation.

---

Input Requirements

Before running the pipeline, you need to edit two files:

1. config.file – contains global settings, tool paths, and parameters.

2. directory.tsv – defines the sample information and file locations.

   The sample directory file (directory.tsv) must have the following 7 columns:

   1. Sample Name
   2. Sample Directory
   3. Normal File 1 (_1.fq or _1.fq.gz)
   4. Normal File 2 (_2.fq or _2.fq.gz)
   5. Tumour File 1 (_1.fq or _1.fq.gz)
   6. Tumour File 2 (_2.fq or _2.fq.gz)
   7. HLA Alleles (if unknown, leave as "")

The pipeline supports one mode ONLY:

• Matched: patient tumour and normal RNA-seq data.
• (Work-in-progress)Tumour-only: patient tumour RNA-seq data.

---

Dependencies

It is recommended to use Conda for environment setup.

Core Tools

• IRFinder – intron retention quantification
• PepQuery 2.0 – peptide validation (requires Java 1.8.0_311)
• NetMHCpan 4.1 – HLA binding predictions
• Optional: arcasHLA – HLA typing

Python Packages

• SciPy v1.14.1
• pyfaidx v0.8.1.3
• Biopython v1.84
• Pysam v0.22.1
• Multiprocessing
  (or install all via environment.yml)

---

Reference Datasets (IRFinder)

• Annotations: GENCODE Human Release 47
• Genome Reference: GRCh38 Primary Assembly

---

Running the Pipeline

1. Modify pipelineformatched.sh to point to your TSV config file and directories.
2. Run:

bash pipelineformatched.sh --config config.file

---

Output

The main output file is generated at:

/outdir/final_output_for_{patient}.csv

This TSV contains the following columns:

• Chr – chromosome
• Start – start coordinate
• End – end coordinate
• gene_id – Ensembl gene ID
• gene_name – gene symbol
• IRratio – intron retention ratio
• no_of_bind – number of predicted HLA binding events
• hla_bind – specific HLA alleles predicted to bind
• peptideSeq – candidate peptide sequence There will also be multiple sub-files and log files, in order to track and for analysis if needed.

---

📌 Notes

• Ensure all paths in your TSV and script are absolute or relative to the working directory.
• HLA alleles can be provided manually or inferred using arcasHLA.
• Tumour-only mode is still experimental and may require manual adjustments.