AMPSeek

Antimicrobial activity, 3D structure, and toxicity prediction from peptide sequences

Contents

1. Description of AMPSeek
2. Installation
3. Quick Start
4. Input data
5. Outputs
6. Usage
7. Container images
8. Benchmarks
9. Troubleshooting
10. Cite
11. License
12. Contact

---

Description of AMPSeek

AMPSeek is a reproducible, containerized workflow that predicts antimicrobial activity, tertiary structure, and toxicity of peptide sequences in one run. It orchestrates:

• AMPlify for broad‑spectrum antimicrobial activity scoring
• LocalColabFold (MMseqs2‑accelerated AlphaFold) for structure prediction
• tAMPer for hemolytic & cytotoxicity prediction (sequence + structure features)
• A compiled HTML report with interactive plots and per‑peptide summaries

Installation

Install Nextflow (via Conda)

We recommend installing Nextflow in a dedicated Conda environment:

conda create -n ampseek -c bioconda nextflow
conda activate ampseek

Install AMPSeek (via Cloning the Repository)

Clone the repository and run with your preferred container profile.

git clone https://github.com/bcgsc/AMPSeek.git
cd AMPSeek

Check if Singularity or Docker is installed properly:

singularity --version
docker --version

Quick Start

Linux:

To run the the pipeline (with default inputs) on Linux choose either Singularity or Docker:

• Singularity:

nextflow AMPSeek.nf -profile singularity

• Docker:

nextflow AMPSeek.nf -profile docker

macOS:

To run the the pipeline (with default inputs) on macOS:

• Apple Silicon (ARM64)

To run AMPSeek on ARM64, you need to Enable QEMU (once):

docker run --privileged --rm tonistiigi/binfmt --install amd64

Then run the Docker profile:

nextflow AMPSeek.nf -profile docker

Emulation is slower; native AMD64 hosts are recommended for large runs.

• Apple Intel (AMD64)

To run AMPSeek on AMD64:

nextflow AMPSeek.nf -profile docker

Input data

• Format: single FASTA file of peptide sequences
• Default dataset: AMPSeek/data/AMPSeek_data_10.fasta (10 peptides)
• Alternative sets: AMPSeek/data/alternative_data/ (20, 100 peptides)

You may:

• Download from the web at runtime:

  nextflow AMPSeek.nf -profile <docker|singularity> --download_from <url>

  This will place the file in AMPSeek/data/ and remove any others there.

• Manually supply FASTA: place exactly one FASTA in AMPSeek/data/, or set a custom path via --data_path.

Keep only the FASTA you intend to process in the input directory, or point the workflow at a specific file with --data_path.

Outputs

Default output directory: output/

• results.html — interactive report with:

  • summary charts (AMPlify & tAMPer distributions, scatter plots)
  • per‑peptide table (activity & toxicity scores, AMPlify attention, interactive 3D structure)
  • CSV export button

  

• foldings/ — LocalColabFold results per peptide:

  • ZIP per peptide (required by tAMPer), containing five PDBs, 3 prediction cycles, Amber relaxation

• tamper_attention/ — per‑peptide self‑attention maps (if enabled)

Intermediate TSV/CSV products from AMPlify/tAMPer may be cleaned after report compilation.

Usage

Common workflow parameters (set on nextflow run AMPSeek.nf ...):

Flag	Type	Default	Description
--data_path	path	AMPSeek/data/	Directory or file path to FASTA input. If a directory is provided, ensure it contains only one FASTA.
--download_from	URL	none	Download a FASTA from the given URL at runtime (replaces other FASTAs in data/).
--output_path	path	output/	Output directory for results.
--output_file	string	results.html	Name of the final HTML report.
--threads	int	30	Max threads/CPUs to use.
--time	duration	Nextflow default	Per‑process timeout (e.g., 1h, 30 min).
--mem	memory	90% of available	Per‑process memory limit (e.g., 8 GB, 32 GB).

Tool‑specific parameters are fixed to the configurations under which AMPlify and tAMPer were developed/tested. Advanced users may modify AMPSeek.nf directly if necessary.

---

Container images

Step	Image	Size	Arch
RUNAMPLIFY	quay.io/biocontainers/amplify:2.0.1--py36hdfd78af_0	748.683 MB	amd64
RUNCOLABFOLD	biohpc/localcolabfold:1.5	13.51 GB	amd64
RUNTAMPER	itsberkeucar/tamper:latest	4.2 GB	amd64
COMPILERESULTS	itsberkeucar/ampseek-visualization:latest	632.04 MB	amd64

If Docker pulls time out on slow links, set a higher pull timeout in nextflow.config, e.g. docker.pullTimeout = '1h'.

Benchmarks

Wall‑clock time and peak RAM on AMD64 with 30 threads; container images pre‑pulled (download time excluded).

Dataset	Size (KB)	# Peptides	Mean length (aa)	Peak RAM	Wall time (hh:mm:ss)
AMPSeek_data_10.fasta	1	10	121.30	6.5 GB	13:03:07
AMPSeek_data_20.fasta	3	20	121.35	7.0 GB	26:14:39
AMPSeek_data_100.fasta	14	100	121.33	-	129:48:00

Total disk required for images on first run: ~19 GB.

Troubleshooting

Symptom	Likely cause	Fix
Pipeline aborts with Error fetching image to cache	Not enough space in Singularity cache	Set a larger cache via NFX_SINGULARITY_CACHEDIR (export before running).
Visualization step fails; multiple TSV files present	Previous run was interrupted before cleanup	Remove intermediate TSVs in the output directory and re‑run.
Running Docker on ARM64 fails	AMD64 images require emulation	Install QEMU with docker run --privileged --rm tonistiigi/binfmt --install amd64.
Unexpected peptides in results	Multiple FASTAs present in input directory	Keep only the intended FASTA in AMPSeek/data/, or use --data_path to a specific file.
Process timeout on HPC	Queue enforces 30–60 min limits	Increase with --time (e.g., --time '2 h').
Out‑of‑memory error	Per‑process memory too low	Increase with --mem (e.g., --mem '16 GB').
Docker pull times out	Slow network	Add docker.pullTimeout = '1h' to nextflow.config profile.
Nothing ran (no AMPlify/ColabFold/tAMPer)	No FASTA found	Place a FASTA in AMPSeek/data/ or pass --data_path.

---

Cite

If you use AMPSeek, please cite:

• AMPlify — Li et al. (2022) BMC Genomics 23:77. doi:10.1186/s12864-022-08310-4 (GitHub)
• ColabFold/LocalColabFold — Mirdita et al. (2022) Nat Methods 19. doi:10.1038/s41592-022-01488-1 (GitHub)
• tAMPer — Ebrahimikondori et al. (2024) Protein Science. doi:10.1002/pro.5076 (GitHub)
• Nextflow — Di Tommaso et al. (2017) Nat Biotechnol 35:316–319. doi:10.1038/nbt.3820 (GitHub)

License

AMPSeek Copyright (c) 2022 British Columbia Cancer Agency Branch. All rights reserved.

AMPSeek is released under the GNU General Public License v3

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 3.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.

For commercial licensing options, please contact Patrick Rebstein ([email protected]).

Contact

For questions, issues, or feature requests, please open a GitHub issue.

Contributors:

• Berke Uçar · Emily Zhang · Ali Salehi · Rene Warren · Inanc Birol