ww-shapemapper

A WDL workflow for running ShapeMapper RNA structure probing analysis in parallel across multiple samples.

Prerequisites

• Cromwell or other WDL-compatible workflow executor
• Docker installation (if using containerized execution)
• ShapeMapper Docker image
• Input FASTQ files from SHAPE-MaP experiments
• Reference sequences in FASTA format

Workflow Overview

This workflow enables parallel processing of multiple RNA samples through the ShapeMapper pipeline. For each sample, it:

1. Processes paired-end reads from modified and untreated conditions
2. Calculates SHAPE reactivity profiles
3. Generates quality control metrics
4. Organizes outputs in sample-specific directories

Input Description

The workflow takes a structured input format using WDL structs. Each sample requires:

struct SampleInfo {
    String name           # Sample identifier
    File target_fa       # Reference sequence in FASTA format
    File modified_r1     # Forward reads from modified sample
    File modified_r2     # Reverse reads from modified sample
    File untreated_r1    # Forward reads from untreated sample
    File untreated_r2    # Reverse reads from untreated sample
}

Optional Parameters

• primers_fa: FASTA file containing primer sequences
• min_depth: Minimum read depth for reporting reactivity (default: 5000)
• is_amplicon: Boolean flag for amplicon sequencing data (default: false)

Runtime Parameters (per task)

• memory_gb: Memory allocation in GB (default: 16)
• cpu: Number of CPU cores (default: 4)

Usage

1. Create an inputs JSON file:

{
  "ShapeMapperAnalysis.samples": [
    {
      "name": "sample1",
      "target_fa": "sample1.fa",
      "modified_r1": "sample1_mod_R1.fq",
      "modified_r2": "sample1_mod_R2.fq",
      "untreated_r1": "sample1_unmod_R1.fq",
      "untreated_r2": "sample1_unmod_R2.fq"
    }
  ],
  "ShapeMapperAnalysis.primers_fa": "primers.fa",
  "ShapeMapperAnalysis.min_depth": 1000,
  "ShapeMapperAnalysis.is_amplicon": true
}

2. Run the workflow:

# Using Cromwell
java -jar cromwell.jar run shapemapper.wdl -i inputs.json

# Using miniwdl
miniwdl run shapemapper.wdl -i inputs.json

Outputs

For each sample, the workflow produces:

• shape_file: SHAPE reactivity profile (.shape format)
• log_file: Execution logs and quality metrics
• output_dir: Directory containing all output files including:
  • Processed alignment files
  • Quality control plots
  • Summary statistics
  • Intermediate files

Docker Requirements

The workflow uses the shapemapper Docker image from the WILDS Docker Library, specifically getwilds/shapemapper:2.3. If you would like to provide your own, update the docker runtime parameter in the WDL with your specific image:

runtime {
    docker: "your-registry/shapemapper:version"
}

Notes

• Input FASTQ files should be quality-checked and adapter-trimmed if necessary
• Ensure sufficient disk space for intermediate files
• Memory and CPU requirements may need adjustment based on RNA length and sequencing depth
• All file paths in the inputs JSON should be absolute or relative to the workflow execution directory

Troubleshooting

Common issues and solutions:

1. Insufficient memory: Increase memory_gb for longer RNA sequences
2. Disk space errors: Adjust disk_size based on input file sizes
3. Missing outputs: Check log files for execution errors
4. Docker errors: Ensure the specified Docker image is accessible

Citation

If you use this workflow, please cite:

• ShapeMapper: Busan S, Weeks KM. Accurate detection of chemical modifications in RNA by mutational profiling (MaP) with ShapeMapper 2. RNA. 2018, 24(2):143-148.