TRAM-seq analysis pipeline

A nextflow pipeline for analyzing TRAM-seq (transcriptome-wide accessibility mapping by sequencing) data to quantify RNA structure and changes across different experimental conditions.

Overview

This pipeline processes TRAM-seq sequencing data to:

• Map paired-end reads from FASTQ files to the reference genome
• Trim adapters and deduplicate reads for quality control
• Generate and process pileup files for mutation detection
• Filter coverage and normalize DMS modification rates
• Compute element-level statistics for genomic regions and transcripts

Pipeline workflow

Main processing flow

flowchart TD
    subgraph "Input Data"
        F1[Control FASTQ<br/>C1, C2, C3]
        F2[DMS FASTQ<br/>D1, D2, D3]
        F3[Stress Control FASTQ<br/>A1, A2, A3]
        F4[Stress DMS FASTQ<br/>AD1, AD2, AD3]
    end
    
    subgraph "Read Mapping"
        M[MAP_READS<br/>Trim, Deduplicate & Align]
    end

    subgraph "Mapped Data"
        A1[Control BAMs<br/>C1, C2, C3]
        A2[DMS BAMs<br/>D1, D2, D3]
        A3[Stress Control BAMs<br/>A1, A2, A3]
        A4[Stress DMS BAMs<br/>AD1, AD2, AD3]
    end
    
    subgraph "Generate raw rates"
        B[EXTRACT_CHROMOSOME_BAM<br/>Split by chromosome]
        C[GENERATE_PILEUP<br/>Mutation detection]
        D[PROCESS_PILEUP<br/>Strand separation]
    end
    
    subgraph "QC & Normalization"
        E[FILTER_COVERAGE<br/>Remove low coverage sites]
        F[NORMALIZE_RATES<br/>DMS vs Control normalization]
    end
    
    subgraph "Harmonize conditions"
        G[COMBINE_CHROMOSOMES<br/>Create genome-wide tracks]
        H[FIND_COMMON_SITES<br/>Identify shared sites]
        I[SET_COMMON_SITES<br/>Apply common filter]
    end
    
    subgraph "Statistical Analysis"
        J[COMPUTE_ELEMENT_STATS<br/>Basic regions: 3'UTR, 5'UTR, CDS]
        K[COMPUTE_ELEMENT_STATS_SPLICED<br/>Transcript-level analysis]
    end

    F1 & F2 & F3 & F4 --> M
    M --> A1 & A2 & A3 & A4
    A1 & A2 & A3 & A4 --> B
    B --> C
    C --> D
    D --> E
    E --> F
    F --> G
    G --> H
    H --> I
    I --> J
    I --> K

    style F1 fill:#EB0C0C
    style F2 fill:#F55F5F
    style F3 fill:#1D05F2
    style F4 fill:#7060F7
    style A1 fill:#EB0C0C
    style A2 fill:#F55F5F
    style A3 fill:#1D05F2
    style A4 fill:#7060F7
    style J fill:#01913B
    style K fill:#02702F

Loading

Quick start

Clone the repository.

git clone https://github.com/yourusername/dms-mapseq-pipeline.git
cd dms-mapseq-pipeline

Set up reference genome files.

bash assets/genomes/download_chromosomes.sh

Prepare ENSEMBL gene annotations.

bash assets/gtf/prepareGTF.sh

Set up STAR index for mapping.

(This command may need to be modified to optimize for read-length used in sequencing library)

bash assets/STARindex/buildIndex.sh

Install the BBMap suite of tools.

bash bbmap/downloadBBMap.sh

Run the pipeline.

nextflow run main.nf -params-file params.yaml

Requirements

Software dependencies

nextflow (>=22.10.0)
cutadapt (>=4.0)
STAR (>=2.7.0)
Java (>=11)
samtools (>=1.15)
bcftools
bedtools
python (>=3.7)

Setup

1. Reference genome files

The pipeline requires individual chromosome FASTA files:

bash assets/genomes/downloadChromosomes.sh

2. BBMap suite

The pipeline uses BBMap tools for read deduplication:

bash bbmap/downloadBBMap.sh

3. STAR genome index

Build a STAR genome index for read alignment:

bash assets/STARindex/buildIndex.sh

Alternatively, specify the path to an existing STAR index in params.yaml:

star_index: "/path/to/your/STAR/index"

4. Input data structure

Option A: Starting from FASTQ files (recommended)

Organize paired-end FASTQ files in the data/fastq/ directory:

data/fastq/
├── C1s_R1.fastq.gz    # Control replicate 1, read 1
├── C1s_R2.fastq.gz    # Control replicate 1, read 2
├── D1s_R1.fastq.gz    # DMS replicate 1, read 1
├── D1s_R2.fastq.gz    # DMS replicate 1, read 2
├── A1s_R1.fastq.gz    # Stress control replicate 1, read 1
├── A1s_R2.fastq.gz    # Stress control replicate 1, read 2
├── AD1s_R1.fastq.gz   # Stress DMS replicate 1, read 1
└── AD1s_R2.fastq.gz   # Stress DMS replicate 1, read 2

The pipeline will automatically:

• Trim adapters and low-quality bases
• Remove PCR duplicates
• Align reads to the reference genome
• Filter for high-quality primary alignments
• Generate BAM files and QC reports

Option B: Starting from pre-existing BAM files

If you already have processed BAM files, you can skip the MAP_READS module by:

1. Placing BAM files in the expected location
2. Modifying the workflow to start from EXTRACT_CHROMOSOME_BAM

Note: Pre-existing BAMs should be:

• Sorted and indexed
• Filtered for primary alignments
• Mapped to the same reference genome used in the pipeline