LAAVA: Long-read Adeno-Associated Virus Analysis

LAAVA is a comprehensive internal pipeline for detecting, characterizing, and visualizing recombinant AAV (rAAV) integration sites using Oxford Nanopore long-read sequencing data.

⚠️ This repository is a summary only. The full pipeline code is not publicly available due to licensing restrictions and internal use.

---

Overview

LAAVA supports scalable rAAV integration analysis from raw FASTQ reads through to annotated visual summaries across multiple gene therapy research projects.

Pipeline Summary

1. Primary alignment to project-specific AAV transgene or plasmid sequences
2. Secondary alignment to full reference genome (e.g., mm39)
3. Assembly of target-mapped reads using Canu
4. Integration site detection and chromosomal distribution analysis
5. Visualization of mapping rates, read characteristics, and coverage profiles

---

My Role

• Designed and implemented the pipeline core using Bash, R, and shell scripts
• Automated dual-alignment strategy with metadata-driven sample handling
• Built modular project structure to support different targets across studies
• Developed Docker-compatible workflow for HPC (LSF) execution
• Generated publication-ready plots and integration summaries in R

---

Tools & Technologies

Tool	Purpose
minimap2	Long-read alignment to target and genome references
samtools	BAM processing and QC
Canu	Local assembly of high-coverage integration loci
R	Summary statistics and ggplot2 visualizations
Docker	Containerization for reproducible, portable execution
LSF	HPC job scheduling

---

Input Files

• Sample metadata CSV — defines sample names, references, and FASTQ paths
• FASTQ files — stored in a central fasta_files/ directory
• Target reference FASTA files — e.g., rAAV, transgene, or plasmid
• Full genome FASTA — for secondary genome-wide integration alignment

---

Output Summary

• BAMs for primary and secondary alignments
• BED files for predicted integration coordinates
• Coverage reports and per-sample summary statistics
• Plots:
  • combined_capture_efficiency_plot.png
  • secondary_alignment_mapping_rates.png
  • combined_coverage_plot.png
  • violin_plot.png (read length distributions)

---

Integration Site Analysis

• Hotspot identification from BED coordinates
• Chromosome-level distribution analysis
• Coverage threshold tracking at 25x, 50x, 100x, etc.
• Comparative read mapping efficiency across multiple samples

---

Interpretation Examples

Observation	Interpretation
High target alignment + low genome mapping	High integration specificity
Widespread genome mapping + hotspots	Off-target insertion sites present
Low target alignment	Poor capture / sample quality

---

Citation

If referencing this pipeline in your work:

Sophia DeGeorgia. LAAVA: Long-read AAV Integration Analysis Pipeline. Internal research tool. 2024.
Tools include: minimap2, samtools, Canu, ggplot2

Also cite key dependencies:

• Li H. Minimap2. Bioinformatics (2018).
• Koren S. et al. Canu. Genome Research (2017).
• Li H. et al. SAMtools. Bioinformatics (2009).

---

Disclaimer

This repository summarizes an internally developed bioinformatics pipeline.
The source code is not open-source and is not available for distribution.

---

Related Projects

• CANVAS-summary — CNV analysis from low-pass WGS