Caretta-shape – A multiple protein structure alignment and feature extraction suite

Caretta is a software-suite to perform multiple protein structure alignment and structure feature extraction.

The older, slower version of Caretta as described in https://doi.org/10.1016/j.csbj.2020.03.011 can be found at https://git.wur.nl/durai001/caretta

Installation

Requirements

Operating system support

1. Linux and Mac

• All capabilities are supported

2. Windows

• The external tool msms is not available in Windows. Due to this:
  • Feature extraction is not available.
  • features argument in caretta-cli must always be run with --only-dssp.
  • caretta-app is not available.

Software

Caretta works with Python 3.9+ Run the following commands to install external dependencies required for feature extraction (Mac and Linux only) This is not required if you are only using the alignment.

conda install -c salilab dssp
conda install -c bioconda msms

Install the command-line interface:

pip install git+https://github.com/TurtleTools/caretta.git

Install both the command-line interface and the web-application (Mac and Linux only):

pip install "caretta[GUI] @ git+https://github.com/TurtleTools/caretta.git"

Environment variables:

export OMP_NUM_THREADS=1 # this should always be 1
export NUMBA_NUM_THREADS=20 # change to required number of threads

Usage

Command-line Usage

caretta-cli input_pdb_folder
# e.g. caretta-cli test_data  

Options:

Usage: caretta-cli [OPTIONS] INPUT_PDB [MODEL_FOLDER]

  Align protein structures using Caretta.

  Writes the resulting sequence alignment and superposed PDB files to
  "caretta_results". Optionally also outputs a set of aligned feature
  matrices, or the python class with intermediate structures made during
  progressive alignment.

Arguments:
  INPUT_PDB       A folder with input protein files  [required]

Options:
  -p FLOAT                        gap open penalty  [default: 1.0]
  -e FLOAT                        gap extend penalty  [default: 0.01]
  -c, --consensus-weight          weight well-aligned segments to reduce gaps
                                  in these areas  [default: True]
  -f, --full                      Use all vs. all pairwise alignment for
                                  distance matrix calculation (much slower)
  -o, --output PATH               folder to store output files  [default:
                                  caretta_results]
  --fasta / --no-fasta            write alignment in FASTA file format
                                  [default: fasta]
  --pdb / --no-pdb                write PDB files superposed according to
                                  alignment  [default: pdb]
  -t, --threads INTEGER           number of threads to use  [default: 4]
  --features                      extract and write aligned features as a
                                  dictionary of NumPy arrays into a pickle
                                  file
  --only-dssp                     extract only DSSP features
  --class                         write StructureMultiple class with
                                  intermediate structures and tree to pickle
                                  file
  --matrix                        write pre-aligned distance matrix and post-
                                  aligned RMSD, coverage and TM-score matrices
                                  to files
  -v, --verbose                   Control verbosity  [default: True]
  --install-completion [bash|zsh|fish|powershell|pwsh]
                                  Install completion for the specified shell.
  --show-completion [bash|zsh|fish|powershell|pwsh]
                                  Show completion for the specified shell, to
                                  copy it or customize the installation.
  --help                          Show this message and exit.

Web-application Usage (Mac and Linux only)

caretta-app <host-ip> <port> 
# e.g. caretta-app localhost 8091

Then go to localhost:8091/caretta in a browser window.

Features

Read features as:

import pickle

with open('caretta_results/result_features.pkl', 'rb') as f:
    protein_ids, features = pickle.load(f)

The features dictionary is a dictionary of NumPy arrays, with the protein_ids defining the order of rows, the columns being the aligned feature values, and the keys as follows:

• dssp_NH_O_1_index, dssp_NH_O_1_energy, dssp_NH_O_2_index, dssp_NH_O_2_energy, dssp_O_NH_1_index, dssp_O_NH_1_energy, dssp_O_NH_2_index, dssp_O_NH_2_energy: hydrogen bonds; e.g. -3,-1.4 means: if this residue is residue i then N-H of I is h-bonded to C=O of I-3 with an electrostatic H-bond energy of -1.4 kcal/mol. There are two columns for each type of H-bond, to allow for bifurcated H-bonds.
• dssp_acc: number of water molecules in contact with this residue *10. or residue water exposed surface in Angstrom^2.
• dssp_alpha: virtual torsion angle (dihedral angle) defined by the four Cα atoms of residues I-1,I,I+1,I+2. Used to define chirality.
• dssp_kappa: virtual bond angle (bend angle) defined by the three Cα atoms of residues I-2,I,I+2. Used to define bend (structure code ‘S’).
• dssp_phi: IUPAC peptide backbone torsion angles.
• dssp_psi: IUPAC peptide backbone torsion angles.
• dssp_tco: cosine of angle between C=O of residue I and C=O of residue I-1. For α-helices, TCO is near +1, for β-sheets TCO is near -1.
• anm_ca: Fluctuations of alpha carbon atoms based on an Anisotropic network model
• anm_cb: Fluctuations of beta carbon atoms based on an Anisotropic network model
• gnm_ca: Fluctuations of alpha carbon atoms based on a Gaussian network model
• gnm_cb: Fluctuations of beta carbon atoms based on a Gaussian network model
• depth_ca: Depths of alpha carbon atoms
• depth_cb: Depths of beta carbon atoms
• depth_mean: Mean depth of residues

Publications

Janani Durairaj, Mehmet Akdel, Dick de Ridder, Aalt DJ can Dijk. "Fast and adaptive protein structure representations for machine learning." Machine Learning for Structural Biology Workshop, NeurIPS 2020 (https://doi.org/10.1101/2021.04.07.438777)

Poster:

Akdel, Mehmet, Janani Durairaj, Dick de Ridder, and Aalt DJ van Dijk. "Caretta-A Multiple Protein Structure Alignment and Feature Extraction Suite." Computational and Structural Biotechnology Journal (2020) . (https://doi.org/10.1016/j.csbj.2020.03.011)