PLAKAT

PLAKAT — Precised Local Alignment marKers Analysis Tool — is a genome-guided workflow for discovering short, species-specific DNA marker regions from bacterial genomes.

The pipeline is designed for closely related bacterial clades where traditional marker genes such as 16S rRNA, rpoB, hsp65, or MLSA panels may not provide enough resolution for species-level or subspecies-level identification.

PLAKAT searches across the whole reference genome, keeps candidate regions that are conserved in the target species, removes regions shared with close relatives, checks for single-copy behavior, maps retained markers back to the reference genome, annotates marker locations, and designs primers for downstream wet-lab validation.

---

Project goals

PLAKAT aims to:

• discover short marker regions that are specific to a target bacterial species;
• screen candidate markers against close relatives and other non-target genomes;
• retain markers that are conserved in the positive group and absent or divergent in the negative group;
• prioritize single-copy, reference-mappable regions;
• annotate marker positions using reference genome features;
• design PCR/Sanger/qPCR-compatible primers for retained markers.

---

Repository structure

PLAKAT/
├── plakat.py
├── make_manifest.py
├── scripts/
│   ├── 00_download_combined.sh
│   ├── 01_slice_genome.sh
│   ├── 02_neg_screen_unique.sh
│   ├── 03_dedup_cluster.sh
│   ├── 04_blast_vs_positive.sh
│   ├── 05_select_core_singlecopy.sh
│   ├── 06_loose_screen.sh
│   ├── 07_prepare_reference.sh
│   ├── 08_blast_vs_ref.sh
│   ├── 09_annotate_with_gff.sh
│   ├── 10_summarize_genes.sh
│   └── 11_design_primers_from_markers.sh
├── README.md
└── .gitignore

Large downloaded genomes, BLAST databases, intermediate files, logs, and final results are intentionally excluded from this repository.

---

Main scripts

make_manifest.py

Builds standardized genome accession lists and manifest tables from NCBI taxon IDs using the NCBI Datasets CLI.

It can produce:

• an accession list in PLAKAT input format;
• a manifest.tsv file mapping assemblies, sequence accessions, roles, strains, organism names, sources, and sequence names;
• automatic |REF tagging for a selected reference assembly.

Example:

python make_manifest.py \
  --taxon 1781 \
  --levels chromosome,complete \
  --source refseq \
  --out-dir manifests/M.marinum \
  --out M.marinum.txt \
  --manifest M.marinum_manifest.tsv

plakat.py

Main controller script for the PLAKAT workflow.

It accepts:

• a positive genome list;
• a reference genome list;
• one or more negative genome lists.

Example:

python plakat.py \
  --positive-list manifests/M.marinum/M.marinum.txt \
  --reference-list manifests/M.marinum/M.marinum.txt \
  --negative-list manifests/M.ulcerans/M.ulcerans.txt manifests/M.chelonae/M.chelonae.txt \
  --params

---

Input list format

PLAKAT accepts accession lists with one sequence per line.

Recommended format:

chr:NC_000000.1@Strain_name
plasmid:NC_000001.1@Strain_name
chr:NC_000002.1@Reference_strain|REF

Notes:

• chr: marks chromosome sequences.
• plasmid: marks plasmid sequences.
• @Strain_name stores the strain label.
• |REF marks the reference sequence used for coordinate mapping.
• Blank lines and lines beginning with # are ignored.

The same species list can be used as both the positive list and reference list if the reference entry is marked with |REF.

---

Workflow summary

PLAKAT currently performs the following major steps:

1. Download combined FASTAs

   Downloads or reuses cached FASTA files for positive, reference, and negative accessions.

2. Slice reference genome

   Splits the reference genome into overlapping windows, typically 500 bp windows with 250 bp step size.

3. Screen against negative genomes

   Removes reference windows that show similarity to non-target genomes.

4. Deduplicate and cluster

   Removes duplicate windows and clusters near-identical candidates.

5. BLAST against positive genomes

   Checks whether candidate markers are present and conserved across target genomes.

6. Select core single-copy candidates

   Keeps markers that satisfy positive-group conservation and chromosome-only single-copy criteria.

7. Loose negative screen

   Applies a relaxed out-group screen to remove borderline cross-reactive markers.

8. Prepare reference bundle

   Builds a reference FASTA and GFF bundle for coordinate mapping and annotation.

9. Map markers back to reference

   BLASTs final markers against the reference genome and creates BED coordinates.

10. Annotate markers

Intersects marker coordinates with GFF features to identify nearby or overlapping genes.

11. Summarize marker genes

Creates a compact marker summary table.

12. Design primers

Uses Primer3 to design primer pairs for retained markers.

---

Main outputs

Typical generated outputs include:

results/sliced/
results/filtered/
results/dedup/
results/blast/
results/core/
results/annotation/
results/primer/
results/logs/
results/params/

Important final outputs may include:

results/core/<target>_final_markers.fasta
results/core/<target>_marker_annotations.tsv
results/core/<target>_marker_summary.tsv
results/primer/<target>_primers.tsv
results/logs/<run_id>/plakat.log
results/params/<target>.yml

These files are generated during analysis and are not tracked by Git.

---

Software requirements

PLAKAT is intended to run in a Unix-like environment such as Linux, macOS, or Windows WSL2.

Core requirements include:

• Python 3
• Bash
• NCBI Datasets CLI
• Entrez Direct / efetch
• BLAST+
• SeqKit
• CD-HIT
• BEDTools
• Primer3
• EMBOSS seqret for some reference-preparation steps

Some optional downstream analyses may require additional tools.

---

Installation notes

Clone the repository:

git clone https://github.com/mingzapingin/PLAKAT.git
cd PLAKAT

Make sure scripts are executable:

chmod +x scripts/*.sh

Check that required tools are available:

which datasets
which efetch
which blastn
which makeblastdb
which seqkit
which cd-hit-est
which bedtools
which primer3_core

---

Example dry run

Use --dry to check input parsing without running the full workflow:

python plakat.py \
  --positive-list path/to/positive.txt \
  --reference-list path/to/reference.txt \
  --negative-list path/to/negative1.txt path/to/negative2.txt \
  --dry

---

Example full run

python plakat.py \
  --positive-list path/to/M.marinum.txt \
  --reference-list path/to/M.marinum.txt \
  --negative-list path/to/M.ulcerans.txt path/to/M.chelonae.txt path/to/M.shottsii.txt \
  --params

Use --verbose to print full tool output to the console:

python plakat.py \
  -p path/to/positive.txt \
  -r path/to/reference.txt \
  -n path/to/negative1.txt path/to/negative2.txt \
  --params \
  --verbose

---

Git tracking policy

This repository tracks only source code and documentation.

The following are intentionally ignored:

• downloaded genomes;
• BLAST databases;
• cached NCBI files;
• intermediate FASTA files;
• result tables;
• logs;
• temporary input lists;
• large analysis outputs.

This keeps the GitHub repository lightweight and reproducible.

---

Author

Kimhun Tuntikawinwong

PLAKAT was developed for genome-guided discovery of species-specific marker regions in high-similarity bacterial clades, with an initial focus on Mycobacterium species.

---

License

License information has not yet been specified.

Before reuse or redistribution, please contact the author.