ScanEPIC

ScanEPIC (Scan Exitron Platform Integration Caller) is a bioinformatics tool for identifying and quantifying exitron splicing events across multiple RNA-seq platforms.

Overview

Exitrons are novel intronic regions with both splice sites contained within a single annotated exon, effectively mimicking genomic deletions. ScanEPIC implements stringent filtering criteria to distinguish genuine exitron splicing events from technical artifacts such as those arising from reverse transcription slippage during library preparation.

Typical runtime is around ~10 minutes using 4 cores on an average sized RNA-seq experiment.

Installation

Prerequisites

• Python 3.9+
• Conda (recommended)

Install from source

Option 1: Simple installation (recommended)

# Create conda environment
conda create -n scanepic python=3.9
conda activate scanepic

# Install build dependencies first
conda install -c conda-forge cython numpy

# Clone and install ScanEPIC
git clone https://github.com/ylab-hi/ScanEPIC.git
cd ScanEPIC
pip install .

Input Requirements

Required Files

• BAM files: Aligned RNA-seq reads with index (.bai)
• Genome FASTA: Reference genome sequence with index (.fai)
• GTF annotation: Gene annotation file (must be bgzip compressed and tabix indexed using htslib package). the gffutils package will create a database file when ScanEPIC runs for the first time. This may take ~20 minutes but will only need to be done once. Example of fully proccessed GTF file of GENCODE v37 can be found here https://doi.org/10.5281/zenodo.15724292.

File Preparation

# If needed, sort your GTF annotation
awk '$1 ~ /^#/ {print $0;next} {print $0 | "sort -k1,1 -k4,4n -k5,5n"}' in.gtf > out_sorted.gtf

# Compress and index GTF file
bgzip annotation.gtf
tabix -p gff annotation.gtf.gz

# Index genome FASTA
samtools faidx genome.fa

# Index BAM file
samtools index input.bam

Quick Start

Test installation

scanepic --help

Short-read RNA-seq

scanepic extract short \
    -i input.bam \
    -g genome.fa \
    -r annotation.gtf.gz \
    -o output.exitron \
    --cores 4

Long-read RNA-seq

scanepic extract long \
    -i input.bam \
    -g genome.fa \
    -r annotation.gtf.gz \
    -o output.exitron \
    --cores 4

Single-cell RNA-seq

NOTE: The scRNA module takes as input a list of BAM files in the following format:

bam	sample_id	group
sc_test_normal.bam	BT1293	g1
sc_test_tumor.bam	BT1301	g2

where the bam column is a path to the BAM file. sample_id will be used to identify sample names. The group column can be used for downstream analysis.

Also required is a TSV file with three columns which identify the cell barcode and its corresponding cell group:

cells	sample_id	cell_group
AACGGTACAAAAGC-1	BT1249	immune
AACTCGGATCGCCT-1	BT1301	cancer
AAGAAGACACTGGT-1	BT1300	blood

Make sure the columns sample_id in both files matches (i.e. cells from sample_id are found in the corresponding BAM file). For examples of these files, see test_data folder. Example command:

scanepic extract single \
    -i bam_list.tsv \
    -t cell_types.tsv \
    -g genome.fa \
    -r annotation.gtf.gz \
    -o output_directory \
    --cores 4

Output Format

ScanEPIC produces tab-separated files containing:

Column	Description
chrom	Chromosome
start	Exitron start position
end	Exitron end position
name	Exitron identifier
region	Genomic region (CDS, 3'UTR, 5'UTR)
ao	Alternative observations (supporting reads)
strand	Strand orientation
gene_symbol	Gene symbol
length	Exitron length
splice_site	Splice site motif
transcript_id	Transcript identifier
pso	Percent spliced-out
dp	Read depth
alignment50	Alignment quality score
rt_repeat	RT-PCR repeat content

We recommend filtering out exitrons with alignment50 score >= 0.7. Exitrons with rt_repeat >= 6 should also be filtered out unless working with direct RNA-seq data.

For scRNA-seq, ScanEPIC produces a folder with tab-separated files for each BAM file containing:

Column	Description
chrom	Chromosome
start	Exitron start position
end	Exitron end position
name	Exitron identifier
region	Genomic region (CDS, 3'UTR, 5'UTR)
ao	Alternative observations (supporting reads)
strand	Strand orientation
gene_symbol	Gene symbol
length	Exitron length
splice_site	Splice site motif
transcript_id	Transcript identifier
pso	Percent spliced-out
dp	Read depth
alignment50	Alignment quality score
rt_repeat	RT-PCR repeat content
read_total	Total number of reads supporting exitron across all cell types
cell_type	Cell type of called exitron
exitron_mols	Total number of unique molecules supporting the exitron (after deduplicating with UMI tags)
unique_mols	Total number of unique molecules at this locus
exitron_cells_not_found	PSO of exitron in cells not in annotated cell type input TSV
pso	PSO of exitron in cell_type (exitron_mols/(unique_mols))
pairwise_z_test_pvals	Comma separated p-values for each cell type, using z-test to compare each PSO pairwise
meta_pval	Fisher meta-p-value of pairwise_z_test_pvals. Good for first glance detection of exitrons that may be dysregulated across cell types

ScanEPIC in scRNA mode also outputs a splicing summary file exitron_splicing_summary.tsv. For each exitron detected in any sample, this file extracts the spliced and non-spliced molecule counts in each sample. This is useful for downstream differential expression analysis (see manuscript).

Currently, ScanEPIC requires reads have the CB and UB tags which are generated with the usual scRNA alignment pipelines.

Test Data

ScanEPIC includes test datasets to verify installation and demonstrate usage. You can download GENCODE v37 annotations that were processed by the instructions above here: https://doi.org/10.5281/zenodo.15724292.

Short-read RNA-seq Test

scanepic extract short \
    -i test_data/test_data.bam \
    -g hg38.fa \
    -r gencode.v37.annotation.sorted.gtf.gz \
    -o test_output_short.exitron \
    --cores 4

Long-read RNA-seq Test

scanepic extract long \
    -i test_data/long_test_data.bam \
    -g hg38.fa \
    -r gencode.v37.annotation.sorted.gtf.gz \
    -o test_output_long.exitron \
    --cores 4

Single-cell RNA-seq Test

scanepic extract single \
    -i test_data/sc_test_bam_list.txt \
    -t test_data/sc_test_cell_types.tsv \
    -g hg38.fa \
    -r gencode.v37.annotation.sorted.gtf.gz \
    -o test_output_sc \
    --cores 4

Citation

If you use ScanEPIC in your research, please cite:

Fry J, Liu Q, Lu X, et al. 3'UTR exitron splicing reshapes the 3' regulatory landscape in cancer. Manuscript in preparation.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Support

For questions, issues, or feature requests:

• Open an issue on GitHub
• Contact: joshua.fry@northwestern.edu