The balance between defence systems and horizontal gene transfer shapes adaptation in clinical strains of Acinetobacter spp.

Published: Muthuraman V, Roy P, Dean P, Lopes BS, Shehreen S. (2026). Journal of Applied Microbiology, lxag069. DOI: 10.1093/jambio/lxag069

---

Overview

Statistical analysis and visualisation repository for the above publication. Investigates the interplay between bacterial defence systems, antibiotic resistance gene carriage, and horizontal gene transfer dynamics across clinical Acinetobacter species.

A. baumannii is a WHO critical-priority pathogen — notorious for acquiring resistance and evading clinical interventions. The relationship between its phage defence architecture and resistance gene carriage was poorly characterised at scale. This repository contains all R and Bash code used to perform that analysis and produce the published figures.

---

Dataset

Two datasets used in the analysis, both available via NCBI:

Dataset	Description	Size
Complete genomes	132 complete Acinetobacter genomes across 18 species (43 A. baumannii, 27 A. pittii, 62 other)	132 genomes
Clinical isolates	90 contig-level assemblies of A. baumannii clinical isolates	90 genomes

NCBI accession numbers for all genomes are listed in data/metadata/accession_list.txt.

---

Repository Structure

Acinetobacter-defence-systems/
├── code/
│   ├── 1_pipeline/               # Bash scripts: genome download through tool execution
│   │   ├── 01_download_genomes.sh
│   │   ├── 02_run_defensefinder.sh
│   │   ├── 03_run_padloc.sh
│   │   ├── 04_run_resfinder.sh
│   │   ├── 05_consolidate_results.sh
│   │   ├── 06_consolidate_resfinder.sh
│   │   ├── 07_run_antidefensefinder.sh
│   │   ├── 08_perform_tblastn.sh
│   │   ├── 09_crispr_pipeline_script.sh
│   │   └── 10_HMRG_perform_tblastn.sh
│   └── 2_analysis/               # R scripts: statistical analysis and figure generation
│       ├── 1_defence_system_analysis.R
│       ├── 2_defence_cooccurrence_analysis.R
│       ├── 3_ic2_clone_analysis.R
│       ├── 4_resistance_gene_analysis.R
│       ├── 5_antidefense_analysis.R
│       ├── 6_ime_analysis.R
│       ├── 7_HMRG_analysis.R
│       ├── 8_final_correlation_analysis.R
│       └── 9_crispr_cas_analysis.R
├── data/
│   └── metadata/
│       ├── accession_list.txt    # NCBI accession numbers for all 222 genomes
│       └── Acinetobacter_metadata.xlsx
├── environment/                  # Conda environment files per tool
│   ├── defensefinder.yaml
│   ├── padloc.yaml
│   ├── resfinder.yaml
│   └── crisprcasfinder_env.yaml
├── methods/
│   └── figures/
│       └── methodology_workflow.png
├── results/
│   └── figures/
│       └── main/                 # Figures 1-5
└── LICENSE

---

Methodology

The pipeline runs in two stages:

Stage 1: Bioinformatic pipeline (code/1_pipeline/)

1. Genome download from NCBI using accession list
2. Defence system prediction with DefenseFinder and PADLOC
3. Antibiotic resistance gene identification with ResFinder
4. Anti-defence system prediction with AntiDefenseFinder
5. Integrative mobile element (IME) identification via tBLASTn against ICEberg
6. CRISPR-Cas system characterisation with CRISPRCasFinder
7. Heavy metal resistance gene (HMRG) identification

Stage 2: Statistical analysis (code/2_analysis/)

R scripts run in numbered order, each producing figures and summary tables for the corresponding section of the paper.

---

Installation

Each bioinformatic tool has its own conda environment to avoid dependency conflicts.

# Clone the repository
git clone https://github.com/vikos77/Acinetobacter-defence-systems.git
cd Acinetobacter-defence-systems

# Create tool environments
conda env create -f environment/defensefinder.yaml
conda env create -f environment/padloc.yaml
conda env create -f environment/resfinder.yaml
conda env create -f environment/crisprcasfinder_env.yaml

R packages (install once in R 4.4.0+):

install.packages(c(
  "tidyverse", "ggplot2", "ggrepel", "ggupset",
  "pheatmap", "viridis", "RColorBrewer", "reshape2",
  "patchwork", "cowplot", "gridExtra", "scales"
))

---

Figures

Figure 1: Epidemiological distribution and genomic characteristics

Geographic distribution of isolates (A), isolation source breakdown — 55% clinical-human (B), GC content vs genome size across species (C), and genomic feature comparison across A. baumannii, A. pittii, and other spp. (D).

---

Figure 2: Defence system distribution and species-specific patterns

Distribution of defence system counts per genome (A), top 20 defence system types across the dataset (B), and species-specific defence system profiles comparing A. baumannii, A. pittii, and others (C).

---

Figure 3: IC2 clone defence system characterisation

Defence system counts in IC2 vs other A. baumannii (Wilcoxon p < 0.0001) (A), prevalence comparison (B), enrichment analysis by odds ratio with FDR correction showing SspBCDE enrichment in IC2 (C), and defence system combinations in IC2 contigs (D).

---

Figure 4: Defence system co-occurrence analysis

Chord diagram of defence system co-occurrence network (A) and co-occurrence matrix showing statistically significant pairings (FDR-corrected, p < 0.05) by odds ratio (B).

---

Figure 5: Comprehensive defence-resistance-HGT correlation analysis

Correlation matrix of defence system categories vs anti-defence, mobile elements, and resistance genes (A), defence system vs IME protein function heatmap (B), and defence system vs resistance gene association heatmap (C).

---

Key Findings

Finding	Detail	Clinical Relevance
IC2 defence specialisation	IC2 clones carry SspBCDE systems (1-2 total vs. ~5 in other strains)	Streamlined architecture facilitates resistance acquisition
Defence-resistance trade-off	RM systems inversely correlate with resistance gene load (r = -0.11)	RM systems restrict horizontal gene transfer
Mobile element facilitation	SspBCDE and Gao_Qat associate positively with HGT	Co-acquisition of defence and resistance via shared genomic neighbourhoods
Species-specific profiles	Distinct defence architectures across 18 Acinetobacter species	Species-targeted therapeutic and phage therapy design

---

Citation

@article{muthuraman2026acinetobacter,
  title={The balance between defence systems and horizontal gene transfer shapes
  adaptation in clinical strains of Acinetobacter spp.},
  author={Muthuraman, Vigneshwaran and Roy, Proyash and Dean, Paul and
  Lopes, Bruno Silvester and Shehreen, Saadlee},
  year={2026},
  journal={Journal of Applied Microbiology},
  pages={lxag069},
  doi={10.1093/jambio/lxag069},
  url={https://doi.org/10.1093/jambio/lxag069}
}

---

Contact

Lead Analyst: Vigneshwaran Muthuraman (vigneshwaran0594@gmail.com) Corresponding Author: Saadlee Shehreen (s.shehreen@tees.ac.uk) Institution: Teesside University, UK