Introduction

nf-core/proteinfold is a bioinformatics best-practice analysis pipeline for Protein 3D structure prediction.

The pipeline is built using Nextflow, a workflow tool to run tasks across multiple compute infrastructures in a very portable manner. It uses Docker/Singularity containers making installation trivial and results highly reproducible. The Nextflow DSL2 implementation of this pipeline uses one container per process which makes it much easier to maintain and update software dependencies. Where possible, these processes have been submitted to and installed from nf-core/modules in order to make them available to all nf-core pipelines, and to everyone within the Nextflow community!

On release, automated continuous integration tests run the pipeline on a full-sized dataset on the AWS cloud infrastructure. This ensures that the pipeline runs on AWS, has sensible resource allocation defaults set to run on real-world datasets, and permits the persistent storage of results to benchmark between pipeline releases and other analysis sources. The results obtained from the full-sized test can be viewed on the nf-core website.

Pipeline summary

1. Choice of protein structure prediction method:

   i. AlphaFold2 - Regular AlphaFold2 (MSA computation and model inference in the same process)

   ii. AlphaFold2 split - AlphaFold2 MSA computation and model inference in separate processes

   iii. ColabFold - MMseqs2 API server followed by ColabFold

   iv. ColabFold - MMseqs2 local search followed by ColabFold

   v. ESMFold - Regular ESM

Usage

[!NOTE] If you are new to Nextflow and nf-core, please refer to this page on how to set-up Nextflow. Make sure to test your setup with -profile test before running the workflow on actual data.

Now, you can run the pipeline using:

nextflow run nf-core/proteinfold \
   -profile <docker/singularity/.../institute> \
   --input samplesheet.csv \
   --outdir <OUTDIR>

The pipeline takes care of downloading the databases and parameters required by AlphaFold2, Colabfold or ESMFold. In case you have already downloaded the required files, you can skip this step by providing the path to the databases using the corresponding parameter [--alphafold2_db], [--colabfold_db] or [--esmfold_db]. Please refer to the usage documentation to check the directory structure you need to provide for each of the databases.

• The typical command to run AlphaFold2 mode is shown below:

  nextflow run nf-core/proteinfold \
    --input samplesheet.csv \
    --outdir <OUTDIR> \
    --mode alphafold2 \
    --alphafold2_db <null (default) | DB_PATH> \
    --full_dbs <true/false> \
    --alphafold2_model_preset monomer \
    --use_gpu <true/false> \
    -profile <docker/singularity/podman/shifter/charliecloud/conda/institute>

• Here is the command to run AlphaFold2 splitting the MSA from the prediction execution:

  nextflow run nf-core/proteinfold \
    --input samplesheet.csv \
    --outdir <OUTDIR> \
    --mode alphafold2 \
    --alphafold2_mode split_msa_prediction \
    --alphafold2_db <null (default) | DB_PATH> \
    --full_dbs <true/false> \
    --alphafold2_model_preset monomer \
    --use_gpu <true/false> \
    -profile <docker/singularity/podman/shifter/charliecloud/conda/institute>

• Below, the command to run colabfold_local mode:

  nextflow run nf-core/proteinfold \
    --input samplesheet.csv \
    --outdir <OUTDIR> \
    --mode colabfold \
    --colabfold_server local \
    --colabfold_db <null (default) | PATH> \
    --num_recycles_colabfold 3 \
    --use_amber <true/false> \
    --colabfold_model_preset "AlphaFold2-ptm" \
    --use_gpu <true/false> \
    --db_load_mode 0
    -profile <docker/singularity/podman/shifter/charliecloud/conda/institute>

• The typical command to run colabfold_webserver mode would be:

  nextflow run nf-core/proteinfold \
    --input samplesheet.csv \
    --outdir <OUTDIR> \
    --mode colabfold \
    --colabfold_server webserver \
    --host_url <custom MMSeqs2 API Server URL> \
    --colabfold_db <null (default) | PATH> \
    --num_recycles_colabfold 3 \
    --use_amber <true/false> \
    --colabfold_model_preset "AlphaFold2-ptm" \
    --use_gpu <true/false> \
    -profile <docker/singularity/podman/shifter/charliecloud/conda/institute>

  [!WARNING]

  If you aim to carry out a large amount of predictions using the colabfold_webserver mode, please setup and use your own custom MMSeqs2 API Server. You can find instructions here.

• The esmfold mode can be run using the command below:

  nextflow run nf-core/proteinfold \
    --input samplesheet.csv \
    --outdir <OUTDIR> \
    --mode esmfold \
    --esmfold_model_preset <monomer/multimer> \
    --esmfold_db <null (default) | PATH> \
    --num_recycles_esmfold 4 \
    --use_gpu <true/false> \
    -profile <docker/singularity/podman/shifter/charliecloud/conda/institute>

[!WARNING] Please provide pipeline parameters via the CLI or Nextflow -params-file option. Custom config files including those provided by the -c Nextflow option can be used to provide any configuration except for parameters; see docs.

For more details and further functionality, please refer to the usage documentation and the parameter documentation.

Pipeline output

To see the results of an example test run with a full size dataset refer to the results tab on the nf-core website pipeline page. For more details about the output files and reports, please refer to the output documentation.

Credits

nf-core/proteinfold was originally written by Athanasios Baltzis (@athbaltzis), Jose Espinosa-Carrasco (@JoseEspinosa), Luisa Santus (@luisas) and Leila Mansouri (@l-mansouri) from The Comparative Bioinformatics Group at The Centre for Genomic Regulation, Spain under the umbrella of the BovReg project and Harshil Patel (@drpatelh) from Seqera Labs, Spain.

Many thanks to others who have helped out and contributed along the way too, including (but not limited to): Norman Goodacre and Waleed Osman from Interline Therapeutics (@interlinetx), Martin Steinegger (@martin-steinegger) and Raoul J.P. Bonnal (@rjpbonnal)

We would also like to thanks to the AWS Open Data Sponsorship Program for generously providing the resources necessary to host the data utilized in the testing, development, and deployment of nf-core proteinfold.

Contributions and Support

If you would like to contribute to this pipeline, please see the contributing guidelines.

For further information or help, don't hesitate to get in touch on the Slack #proteinfold channel (you can join with this invite).

Citations

If you use nf-core/proteinfold for your analysis, please cite it using the following doi: 10.5281/zenodo.7437038

An extensive list of references for the tools used by the pipeline can be found in the CITATIONS.md file.

You can cite the nf-core publication as follows:

The nf-core framework for community-curated bioinformatics pipelines.

Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen.

Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.