Here's a README.md draft for your GitHub repository that describes the analysis pipeline for PTA and WGS data:

---

PTA-WGS Analysis Pipeline

This repository contains a workflow description and tools for analyzing Primary Template-directed Amplification (PTA) and Whole Genome Sequencing (WGS) data. The pipeline includes all steps necessary for high-quality variant detection, including mapping, duplicate marking, realignment, variant calling, and filtration.

Overview

This pipeline processes WGS reads mapped against the GRCh38 reference genome, focusing on:

• Handling PTA-specific artifacts using advanced machine learning models (PTATO Random Forest Model).
• High-quality variant calling with GATK.
• Comprehensive variant filtration for somatic and germline mutation detection.

Key Features

• Support for both FASTQ and CRAM inputs.
• Modular WDL workflow for running on Terra or locally.
• Incorporation of existing GATK tools and custom PTA-specific artifact handling.

---

Workflow Steps

1. Input Files

• FASTQ or CRAM files: Raw sequencing data.
• Reference Genome: GRCh38.
• Additional inputs: BED files, BAM indices, and genome-specific annotations.

2. Pipeline Steps

Step 1: Mapping Reads (if FASTQ provided)

• Tool: BWA mem
• Output: SAM/BAM files.

Step 2: Processing CRAM/BAM Files

• Duplicate marking using Sambamba.
• Realignment with GATK BaseRecalibrator.

Step 3: Variant Calling

• Tool: GATK HaplotypeCaller
• Multi-sample mode with EMIT_ALL_CONFIDENT_SITES.

Step 4: Variant Filtration

• Tool: GATK VariantFiltration
• Filters:
  • QD < 2.0
  • MQ < 40.0
  • FS > 60.0
  • HaplotypeScore > 13.0
  • SOR > 4.0 and more.

Step 5: PTA Artifact Filtering

• PTATO Random Forest Model trained with 26 genomic features:
  • Allelic imbalance (most important).
  • DNA replication timing.
  • Distance to the nearest gene.
  • Repeat regions and sequence context.

---

Installation and Dependencies

Requirements

• Docker
• Cromwell/WDL runtime (for Terra)
• Tools included:
  • BWA (v0.7.17)
  • Sambamba (v0.6.8)
  • GATK (v4.1.3.0)
  • Samtools (v1.10)

Installation

1. Clone the repository:

   git clone [email protected]:broadinstitute/PTA_Analysis.git
   cd PTA_Analysis

2. Install dependencies:

   • Use Docker containers for all tools.

3. Prepare input files (FASTQ or CRAM, Reference Genome).

---

Running the Pipeline

On Terra

1. Import the WDL workflows into Terra.
2. Set up input JSON files for each step.

Locally

Run the pipeline using Cromwell:

java -jar cromwell.jar run main.wdl -i inputs.json

---

Repository Contents

• workflows/: WDL workflows for each pipeline step.
• tasks/: Individual WDL tasks for mapping, duplicate marking, realignment, and variant calling.
• docker/: Dockerfiles for each tool used in the pipeline.
• examples/: Example input files and configuration JSONs.

---

Resources

• GATK Documentation: https://gatk.broadinstitute.org/
• PTATO Tool: https://github.com/ToolsVanBox/PTATO
• Reference Genome (GRCh38): Ensembl