AlphaFold-based Protein Analysis Pipeline

---

📝 Table of Contents

• Description
• Installation
• Usage
• Authors
• Acknowledgments

Description

This project constitutes a protein analysis pipeline that allows for a quick and comprehensive analysis of a protein sequence or structure.

• The pipeline accepts as input a protein sequence in FASTA format or a protein structure in PDB format. If a PDB file is not provided, the 3D structure of the protein can be predicted using AlphaFold2.

• A sequence-based analysis can be performed, including determining physicochemical properties and aligning the protein sequence against other databases such as Pfam and Swiss-Prot/UniRef90.

• A structure-based analysis can be performed, including predicting the effect of amino acid substitutions over the protein stability using SimBa2 and the detection of binding pockets using P2Rank.

• A list of ligands can be specified in the PDB/MOL2 format. The binding affinity of the protein-ligand interactions can be predicted using AutoDock Vina.

• The outputs obtained during each process are aggregated into a MultiQC HTML report. The analysis report presents the results in an interactive manner, including visualizing the three-dimensional protein structure using the iCn3D web viewer. The pipeline is being developed using Nextflow.

For upcomming updates and ideas see our Roadmap.

Installation

IMPORTANT: This project is under active development. Momentarily the pipeline can only be utilized by manually installing the desired packages and tools. Support for Docker/Singularity containers represents a high-priority future update.

Prerequisites:

• Python ≥ 3.8
• Java ≥ 11
• git, pip, conda/mamba

Mandatory installation:

• The minimum setup required for running the pipeline. This configuration allows for executing the sequence properties and 3D structure viewer components.

conda config --env --add channels conda-forge
conda config --env --add channels anaconda
conda config --env --add channels bioconda

conda install 'numpy>=1.18.5' 'pandas>=1.4.1' 'biopython>=1.76' 'multiqc>=1.12' pymol-open-source=2.5.0 graphviz
conda install -c salilab dssp=3.0.0

• Install Nextflow and add the executable to PATH: https://github.com/nextflow-io/nextflow

curl -fsSL https://get.nextflow.io | bash

• Clone this repository:

git clone https://github.com/OtimusOne/AFPAP.git

Optional components:

The following analysis steps are optional and their installation can be skipped if so desired.

3D structure prediction - AlphaFold2/ColabFold:

• AlphaFold2 is used to predicted the protein structure if a PDB file is not provided. We use a local instance of ColabFold in order to avoid the large databases used by native AlphaFold2 implementation.
• Follow the install intructions at: https://github.com/YoshitakaMo/localcolabfold
• If ColabFold is not installed the user must provide a PDB structure using the --pdb argument if structural analysis is desired. Set skipAlphaFold = true inside nextflow.config.

Pfam search:

• The Pfam database is used to match the protein sequence agains protein families.
• Download the Pfam database(~1.5GB) - Pfam-A.hmm Pfam-A.hmm.dat: http://ftp.ebi.ac.uk/pub/databases/Pfam/current_release/
• Install the following packages:

conda install pfam_scan perl-json

• Inside the database directory execute:

hmmpress Pfam-A.hmm

• Set path variable inside nextflow.config

params {
    ...
    pfam_path="/path/to/Pfam/directory"
    ...
}

• If Pfam is not used set skipPfamSearch = true inside nextflow.config.

Conservation MSA:

• The protein sequence conservation can be computed by generating a Multiple Sequence Alignment using a sequence database.
• Install Blast, Muscle and CD-HIT:

conda install blast muscle cd-hit

• Create a BLAST database from which to generate the MSA file. You can use any FASTA database such as SwissProt/TrEMBL/UniRef50/UniRef90 (https://ftp.uniprot.org/pub/databases/uniprot/current_release/).

  Example setting UniRef90 up as a Blast database: https://ftp.uniprot.org/pub/databases/uniprot/current_release/uniref/uniref90/

curl {https://ftp.uniprot.org/pub/databases/uniprot/current_release/uniref/uniref90/uniref90.fasta.gz} | gunzip | makeblastdb -out uniref90 -dbtype prot -title UniRef90 -parse_seqids

• Set path variable inside nextflow.config

params {
    ...
    blast_path="/path/to/BLASTDB/uniref90"
    ...
}

• warning: SwissProt might not return enough hits to calculate the sequence conservation, downloading large databases such as TrEMBL and UniRef90 might take a lot of time
• If this component is not used set skipConservationMSA = true inside nextflow.config.

Protein stability changes - SimBa2:

• SimBa2 is used to predict the effect of amino-acid substitutions on protein stability.

python -m pip install https://github.com/kasperplaneta/SimBa2/archive/main.tar.gz

• If SimBa2 is not installed set skipStabilityChanges = true inside nextflow.config.

Pocket prediction - P2Rank:

• P2Rank is used to predict ligand-binding pockets from the protein structure.
• Follow the install intructions at: https://github.com/rdk/p2rank
• Pillow is required for generating the pocket visualizations:

conda install pillow

• If P2Rank is not installed set skipPocketPrediction = true inside nextflow.config.

Molecular Docking - AutoDock Vina:

• AutoDock Vina is used to dock the ligands provided with the --ligands argument.
• Install ADFR suite - MSMS not required https://ccsb.scripps.edu/adfr/downloads/
• Install Vina python bindings:

pip install vina

• If the pocket prediction step has been executed beforehand, Vina will dock the ligands against each of the predicted pockets, otherwise it will only execute a blind docking step.
• If Vina is not installed set skipMolecularDocking = true inside nextflow.config.

Usage

Usage example:

nextflow run main.nf --fasta input.fasta
nextflow run path/to/main.nf --pdb input.pdb --ligands "ligand1.mol2 path/to/ligand2.pdb"

For a full list of parameters run:

nextflow run main.nf --help

Authors

• Maghiar Octavian-Florin

Acknowledgements

If you find this work useful please properly cite any of the revelant tools.

• AlphaFold2: Jumper, J. et al. (2021) ‘Highly accurate protein structure prediction with AlphaFold’, Nature, 596(7873), pp. 583–589. Available at: https://doi.org/10.1038/s41586-021-03819-2.

• ColabFold: Mirdita, M. et al. (2021) ColabFold - Making protein folding accessible to all. preprint. Bioinformatics. Available at: https://doi.org/10.1101/2021.08.15.456425.

• P2Rank: Krivák, R. and Hoksza, D. (2018) ‘P2Rank: machine learning based tool for rapid and accurate prediction of ligand binding sites from protein structure’, Journal of Cheminformatics, 10(1), p. 39. Available at: https://doi.org/10.1186/s13321-018-0285-8.

  Jendele, L. et al. (2019) ‘PrankWeb: a web server for ligand binding site prediction and visualization’, Nucleic Acids Research, 47(W1), pp. W345–W349. Available at: https://doi.org/10.1093/nar/gkz424.

• AutoDock Vina: Eberhardt, J. et al. (2021) ‘AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings’, Journal of Chemical Information and Modeling, 61(8), pp. 3891–3898. Available at: https://doi.org/10.1021/acs.jcim.1c00203.

  Trott, O. and Olson, A.J. (2009) ‘AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading’, Journal of Computational Chemistry, p. NA-NA. Available at: https://doi.org/10.1002/jcc.21334.

• SimBa2: Bæk, K.T. and Kepp, K.P. (2022) ‘Data set and fitting dependencies when estimating protein mutant stability: Toward simple, balanced, and interpretable models’, Journal of Computational Chemistry, 43(8), pp. 504–518. Available at: https://doi.org/10.1002/jcc.26810.

• iCn3D: Wang, J. et al. (2020) ‘iCn3D, a web-based 3D viewer for sharing 1D/2D/3D representations of biomolecular structures’, Bioinformatics. Edited by A. Valencia, 36(1), pp. 131–135. Available at: https://doi.org/10.1093/bioinformatics/btz502.

• Nextflow: Di Tommaso, P. et al. (2017) ‘Nextflow enables reproducible computational workflows’, Nature Biotechnology, 35(4), pp. 316–319. Available at: https://doi.org/10.1038/nbt.3820.

• MultiQC: Ewels, P. et al. (2016) ‘MultiQC: summarize analysis results for multiple tools and samples in a single report’, Bioinformatics, 32(19), pp. 3047–3048. Available at: https://doi.org/10.1093/bioinformatics/btw354.

• BLAST: Camacho, C. et al. (2009) ‘BLAST+: architecture and applications’, BMC Bioinformatics, 10(1), p. 421. Available at: https://doi.org/10.1186/1471-2105-10-421.

• CD-HIT: Fu, L. et al. (2012) ‘CD-HIT: accelerated for clustering the next-generation sequencing data’, Bioinformatics, 28(23), pp. 3150–3152. Available at: https://doi.org/10.1093/bioinformatics/bts565.

• MUSCLE: Edgar, R.C. (2004) ‘MUSCLE: multiple sequence alignment with high accuracy and high throughput’, Nucleic Acids Research, 32(5), pp. 1792–1797. Available at: https://doi.org/10.1093/nar/gkh340.