nf-core/proteinfold

Protein 3D structure prediction pipeline

alphafold2colabfoldesmfoldprotein-fold-predictionprotein-foldingprotein-sequencesprotein-structure

Launch version 2.0.0 https://github.com/nf-core/proteinfold

Version history

Download .zip Download .tar.gz View on GitHub

[2.0.0] - 2026-03-27

Enhancements & fixes

[#177] - Fix typo in some instances of model preset alphafold2_ptm.
[PR #178] - Enable running multiple modes in parallel.
[#179] - Produce an interactive html report for the predicted structures.
[#180] - Implement Foldseek.
[#188] - Fix colabfold image to run in gpus.
[PR ##205] - Change input schema from sequence,fasta to id,fasta.
[PR #210] - Moving post-processing logic to a subworkflow, change wave images pointing to oras to point to https and refactor module to match nf-core folder structure.
[#214] - Fix colabfold image to run in cpus after #188 fix.
[PR ##220] - Add RoseTTAFold-All-Atom module.
[PR ##223] - Add HelixFold3 module.
[#235] - Update samplesheet to new version (switch from sequence column to id).
[#239] - Update alphafold2 standard mode Dockerfile.
[#240] - Separate download and input of pdb mmcif files and obsolete database.
[#229] - Add Boltz pipeline PR #227.
[PR #249] - Update pipeline template to nf-core/tools 3.2.0.
[PR #271] - Update RFAA and HF3 dockerfiles for quicker building and reduction in image size.
[PR #274] - Simplify run_helixfold3 module and move arguments to modules.config.
[#276] - Update helixfold3 dockerfile to make it compatible with H100 gpus.
[#259] - Fix esmfold docker image to make it compatible with hopper GPU architecture.
[#281] - Fix how argument --nv is passed to apptainer and singularity in the config.
[PR #283] - Fixes to meet language server requirements and update link of the helixfold3 image.
[PR #287] - Fixes symlinking of every mmcif file causing excess I/O.
[#293] - Add back alphafold2_model_preset input to the call to run_alphafold2_pred.
[PR #294] - Temporary downgrade of schema for passing CI tests with Nextflow edge version.
[#272] - Colouring scheme conforming to AlphaFold2 confidence bands in html report.
[PR #297] - Update pipeline template to nf-core/tools 3.2.1.
[#273] - Fixes comparison report to correctly label msa coverage plots with corresponding method label.
[#290] - Update Alphafold2 split images to make them compatible Hopper gpus.
[PR #302] - Fix HF3 dbs and max_template_date.
[PR #305] - Stop RFAA and HF3 symlinking scripts into workdir.
[PR #306] - extract_output.py -> extract_metrics.py so pLDDT, MSA, PAE emitted as raw data .tsv files
[PR #307] - Update Boltz-1 boilerplate and formatting.
[PR #314] - Fix extract metrics for broken modules.
[PR #312] - pTM & ipTM metrics now extracted
[PR #315] - Add global db flag.
[#263] - Removed broken colabfold options (auto and alphafold2)
[PR #316] - Add process_gpu label to modules which use GPU.
[PR #319] - Update boltz workflow to accept YAML as input.
[PR #322] - Updates and reorganises the reference database directory structure.
[PR #329] - Updates Boltz module to include Boltz-2.
[PR #332] - Fix rare superposition bug in reports.
[PR #333] - Updates the RFAA dockerfile for better versioning and smaller image size.
[PR #335] - Update pipeline template to nf-core/tools 3.3.1.
[PR #346] - Update pipeline template to nf-core/tools 3.3.2.
[PR #351] - add chain-wise (i)pTM values and summary file for AF3-generation codes.
[PR #354] - Update documentation with mode specific usage, capabilities.
[PR #355] - Remove unneccesary params from Boltz and Helixfold3 modes.
[PR #356] - Update AF2 defaults to use split mode and monomer_ptm model.
[PR #357] - Update ColabFold module and image.
[PR #359] - Harmonize parameters across modes.
[PR #360] - Rename some DBs paths in the run modules so they are equal to those when DBs are downloaded.
[PR #362] - Update boltz Dockerfile and image pinning specific version (2.0.3).
[#364] - Move Dockerfiles to its corresponding module.
[PR #370] - Fix extract chain metrics.
[#367] - Boltz post-processing crashes.
[#368] - Helixfold3 iPTM output missing when dealing with monomers make the process to fail.
[#369] - Download all Alphafold3 DBs.
[PR #350] - PAE of model 0 in Boltz HTML report, AlphaFold2 to pass the build system
[PR #377] - Fix sequence msa synch for af2 split.
[#380] - Fixes alphafold2_model_preset bug on retry.
[#382] - Readds --full_dbs as a global option.
[#378] - Fix nested obsolete pdbs from pdb70.
[#388] - Fix colabfold prefix handling for output metrics.
[#387] - Fix alphafold2_standard obsolete.dat path error.
[#389] - Locked version numbers for HelixFold3 image to prevent bug caused by newer mamba versions.
[PR #397] - Fix AF2 mgnify handling and improve version reporting for AlphaFold2 containers.
[PR #398] - Fix issues with PREPARE_DBS subworkflows.
[PR #399] - Update alphafold2 and alphafold2_pred Dockerfiles.
[PR #404] - Boltz cache files moved to workdir, fixed version checks and Boltz stubRun.
[#401] - Get rid of symlinking in the prediction tools processes when using “PREPARE_DBS” subworkflows
[#410] - Switch RosettaFold2NA to Boltz-style multi-chain FASTA inputs and drop the interactions sheet.
[PR #407] - Several changes to meet nf-core standards.
[PR #409] - Force single pdb workflow outputs to return as a list
[PR #396] - Split ColabFold into separate optimised containers with version pinning and significant size reduction.
[#412] - Substitute ”/” with ”_” from fasta headers used to name files when using “—split_fasta”.
[PR #424] - Bump docker image version for release to 2.0.0, make code more friendly with Nextflow language server and other format issues/fixes.
[#423] - Generate json workflow using bioflow-insight.
[#425] - Pass as a single input channel fasta and features to get rid of meta2 in RUN_ALPHAFOLD2_PRED.
[#440] - Support single-letter RF2NA type tags (type=P/R/D/S) in ROSETTAFOLD2NA FASTA headers.
[PR #442] - Bump version 2.6.1 of nf-schema, Nextflow minimum version to 25.10.2 and update utils_nfschema_plugin subworkflow.
[PR #443] - Add documentation guide for contributing new prediction modes.
[PR #446] - Fix warnings from Nextflow lint.
[PR #451] - Remove af2 multimer padding from msa plots.
[#417] - Add boltz_use_kernels parameter to enable/disable using optimized Triton-based CUDA kernels CUDA kernels for Boltz inference.
[#417] - Handle incompatible CUDA kernel errors in Boltz by automatically retrying with --use_kernels false.
[#285] - Adding contributors to manifest.
[PR #460] - Use nvidia-smi to obtain number of SM.
[PR #454] - Update publishdir patterns for alphafold2 modules.
[PR #458] - Update publishdir patterns for colabfold module.
[#313] - Harmonize colabfold metrics extraction with other modes.
[#455] - Fix colabfold monomer inheriting id from fasta header.
[#457] - Fix colabfold multimer always downloading model weights.
[PR #461] - Update publishdir patterns for HelixFold3 module
[PR #462] - Update publishdir patterns for RoseTTAFold-All-Atom modules
[PR #464] - Update publishdir patterns for Boltz module
[PR #466] - Update module conf and publishdir patterns for ESMFold, pass through container args
[PR #469] - HTML reports now in /reports output directory
[PR #468] - Update publishdir patterns for Alphafold3 module
[PR #471] - Update publishdir patterns for Rosettafold2na module
[#473] - Add nf-test for rosettafold-aa, rosettafold2na, helixfold3 and boltz modes.
[PR #475] - Update and simplify outputs.md with the latest structure
[#480] - Make version reporting consistent for all local modules.
[PR #482] - Update utils_nfschema to fix help message with strict syntax.
[PR #483] - Move foldseek logic to the post_processing subworkflow and set sensible time to aria2 processes.
[PR #492] - Clean TODOs from code and create issues instead for 2.0.0 release preparation.
[PR #493] - Standardise Dockerfiles labels and bump version 2.0.0 to prepare release.
[#494] - Publish Colabfold DBs when downloaded to be directly consumable using colabfold_db parameter.
[#496] - Publish all DBs when downloaded to be directly consumable using the corresponding mode parameter.
[#494] - Publish Colabfold DBs when downloaded to be directly consumable using colabfold_db parameter.
[#499] - Get rid of ENTRYPOINT in alphafold2 dockerfiles.
[PR #501] - Move python code of BOLTZ_FASTA to a python script in bin.
[#503] - Add checkIfExists validation to user-provided database paths across all prepare DB subworkflows.
[#507] - Implement missing full tests and check that the others work before release 2.0.0.
[PR #509] - Setup gpu environment for AWS full tests.
[#519] - Fix AWS full test before release 2.0.0.
[PR #525] - Reduce AlphaFold2 multimer full test to a single replicate.
[PR #531] - Fix alphafold2_random_seed type.
[PR #538] - Bump version 2.0.0 for release preparation.
[PR #563] - Fix boltz csv sync.
[PR #570] - Update parameters table in changelog.

Download .zip Download .tar.gz View on GitHub

[1.1.1] - 2024-07-30

Enhancements & fixes

Minor patch release to fix multiqc report.

Download .zip Download .tar.gz View on GitHub

[1.1.0] - 2024-06-25

Credits

Special thanks to the following for their contributions to the release:

Thank you to everyone else who has contributed by reporting bugs, enhancements or in any other way, shape or form.

Enhancements & fixes

[#80] - Add accelerator directive to GPU processes when params.use_gpu is true.
[#81] - Support multiline fasta for colabfold multimer predictions.
[#89] - Fix issue with excessive symlinking in the pdb_mmcif database.
[PR #91] - Update ColabFold version to 1.5.2 and AlphaFold version to 2.3.2
[PR #92] - Add ESMFold workflow to the pipeline.
Update metro map to include ESMFold workflow.
Update modules to remove quay from container url.
[nf-core/tools#2286] - Set default container registry outside profile scope.
[PR #97] - Fix issue with uniref30 missing path when using the full BFD database in AlphaFold.
[PR #100] - Update containers for AlphaFold2 and ColabFold local modules.
[PR #105] - Update COLABFOLD_BATCH docker container, metro map figure and nextflow schema description.
[PR #106] - Add singularity.registry = 'quay.io' and bump NF version to 23.04.0
[#108] - Fix gunzip error when providing too many files when downloading PDBMMCIF database.
[PR #111] - Update pipeline template to nf-core/tools 2.9.
[PR #112] - Use nf-validation plugin for parameter and samplesheet validation.
[#113] - Include esmfold dbs for full data sets.
[PR #114] - Update paths to test dbs.
[PR #117] - Update pipeline template to nf-core/tools 2.10.
[PR #132] - Remove lib/ directory.
[#135] - Reduce Alphafold Docker images sizes.
[#115] - Throw message error when profile conda is used.
[#131] - Add esmfold small tests.
[#144] - Force value channels when providing dbs (downloaded) in main.nf to enable the processing of multiple samples.
[#147] - Update modules to last version.
[#145] - Implement test to check the processes/subworkflows triggered when downloading the databases.
[#130] - Add --skip_multiqc parameter.
[PR #154] - Update pipeline template to nf-core/tools 2.14.1.
[#148] - Update Colabfold DBs.
[PR #159] - Update mgnify paths to new available version.
[PR ##163] - Fix full test CI.
[#150]] - Add thanks to the AWS Open Data Sponsorship program in README.md.
[PR ##166] - Create 2 different parameters for Colabfold and ESMfold number of recycles.