PDBCleanup

Description

PDBCleanup is an R package which offers utilities for improving the quality of PDB protein structure files by selecting high resolution regions and improving relative domain positions by superimposing individual regions against corresponding regions of another reference protein structure. It further allows visualization of a quality score along an input protein’s sequence, visualization of the 3D protein structure produced after selecting high quality regions and performing superimposition, and visualization of the structural difference between a pair of protein structures.

This package is beneficial as a pre-processing step before performing analysis using protein structural data. It is compatible with other common structural biology packages but provides new highly customizable functions to streamline the first steps of structural data preparation.

PDBCleanup was designed with AlphaFold predicted protein structures in mind, which often contain poorly predicted regions (indicated by low pLLDT scores in the B-factor column) or misaligned domains. However, it can be used for any protein structures, including experimental protein structures.

PDBCleanup was developed in an environment using R version 4.3.1 on an aarch64-apple-darwin20 (64-bit) platform running macOS Sonoma 14.0.

Installation

You can install the development version of PDBCleanup from GitHub with:

# install.packages("devtools")
library("devtools")
devtools::install_github("isaiahah/PDBCleanup", build_vignettes = TRUE)
library("PDBCleanup")

To run the Shiny app:

runShiny()

Overview

ls("package:PDBCleanup")
data(package = "PDBCleanup")  # No R datasets are available with the package
browseVignettes("PDBCleanup")

There are 7 functions in PDBCleanup:

1. selectResno: A small function to select a provided set of residues from a PDB protein structure.

2. selectHighResolution: A function to select high resolution regions in a protein structure. High resolution regions are selected using the values in the B-factor column: for experimental structures, this is B-factor which corresponds to uncertainty in position; for AlphaFold predicted structures, this is pLDDT which corresponds loosly to prediction confidence. The user provides a structure loaded using bio3d and whether the structure is an AlphaFold predicted structure or an experimental structure. They can optionally provide a threshold value, otherwise the function uses a default threshold. The function returns a structure compatible with bio3d functions for later analysis.

3. selectOrdered: A function to select the intrinsically ordered regions in a protein structure and remove intrinsically disordered regions. Order is predicted using the FoldIndex score. The user provides a structure loaded using bio3d and optionally provides a window size used to compute FoldIndex score and a threshold used to define ordered regions. The function returns a structure compatible with bio3d functions for later analysis.

4. plotProteinQuality: A function to visualize the B-factor of a bio3d protein structure as a bar plot along the sequence. The user specifies whether the structure is predicted or experimental to specify the y-axis label and gives a custom plot title.

5. alignDomainsRigid: A function to align a mobile protein structure against a fixed template protein structure by independently rigidly aligning each pair of provided domains, with the domain as one unit. The locations of atoms in the mobile protein structure which are not part of a domain do not move. It both accepts and returns protein structures compatible with bio3d functions.

6. alignDomainsSmooth: A variant of alignDomainsRigid which rigidly aligns each domain then smooths the location of each atom based on the displacement of nearby atoms, as described in Terwilliger et al. (2023). Like the above function, it both accepts and returns protein structures compatible with bio3d functions.

7. viewStructure: A function to visualize the provided structure with a 3D cartoon model. This function opens a browser window with the interactive 3D protein model.

The package provides two data files containing protein structures for E. coli protein Pqql, a probable zinc protease with Uniprot ID P31828. These are available for vignettes and represent a case of two very different structures for the same protein, where preprocessing may be needed.

1. 6ofs_experimental.pdb contains an experimental structure uploaded to the PDB under ID 6ofs as part of the publication “Protase-associated import systems are widespread in Gram-negative bacteria” by Grinter R et al (2019). This file has been modified to remove HETATM and ANISOU lines from the file to reduce size.

2. 6ofs_predicted.pdb contains a computationally predicted structure by AlphaFold2 uploaded to the AlphaFold database.

Below is a flowchart describing data preparation possible with this package.

See help(package = "PDBCleanup") for further details and references provided by citation("PDBCleanup").

Contributions

This package was created by Isaiah Hazelwood. The selectResno, selectHighResolution, and selectOrdered functions use the atom.select and trim functions from bio3d. The alignDomainsRigid and alignDomainsSmooth functions use these bio3d functions alongside fit.xyz to perform domain alignments. The selectOrdered function uses the foldIndexR function from idpr. The plotProteinQuality function uses ggplot2 to create plots. The viewStructure function uses r3dmol to create and visualize a 3D model of the protein structure. The author wrote all package functions referencing only documentation and Anjali Silva’s testing package. No generative AI tools were used in the development of this package.

Acknowledgements

This package was developed as part of an assessment for 2023 BCB410H: Applied Bioinformatics course at the University of Toronto, Toronto, Canada. PDBCleanup welcomes issues, enhancement requests, and other contributions. To submit an issue, use the GitHub issues.

References

Chang W, Cheng J, Allaire J, Sievert C, Schloerke B, Xie Y, Allen J, McPherson J, Dipert A, Borges B (2023). shiny: Web Application Framework for R. R package version 1.8.0, https://CRAN.R-project.org/package=shiny.

Grant, B.J. et al. (2006) Bioinformatics 22, 2695–2696.

Grinter R, Leung PM, Wijeyewickrema LC, Littler D, Beckham S, Pike RN, et al. (2019) Protease-associated import systems are widespread in Gram-negative bacteria. PLoS Genet 15(10): e1008435. https://doi.org/10.1371/journal.pgen.1008435

H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York, 2016.

Jumper, J., Evans, R., Pritzel, A. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). https://doi.org/10.1038/s41586-021-03819-2

McFadden, W. M., and Yanowitz, J. L. (2022). idpr: A package for profiling and analyzing Intrinsically Disordered Proteins in R. PLOS ONE, 17(4), e0266929. doi:10.1371/journal.pone.0266929.

Prilusky, J., Felder, C. E., et al. (2005). FoldIndex: a simple tool to predict whether a given protein sequence is intrinsically unfolded. Bioinformatics, 21(16), 3435-3438.

Silva, A. (2022) TestingPackage: An Example R Package For BCB410H. Unpublished. URL https://github.com/anjalisilva/TestingPackage.

Su W, Johnston B (2021). r3dmol: Create Interactive 3D Visualizations of Molecular Data. R package version 0.1.2, https://CRAN.R-project.org/package=r3dmol.

Terwilliger, T.C., Liebschner, D., Croll, T.I. et al. AlphaFold predictions are valuable hypotheses and accelerate but do not replace experimental structure determination. Nat Methods (2023). https://doi.org./10.1038/s41592-023-02087-4

Varadi M., Anyango S., Deshpande M., et al. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Research, Volume 50, Issue D1, 7 January 2022, Pages D439–D444. https://doi.org/10.1093/nar/gkab1061

Zhoutong Sun, Qian Liu, Ge Qu, Yan Feng, and Manfred T. Reetz. Utility of B-Factors in Protein Science: Interpreting Rigidity, Flexibility, and Internal Motion and Engineering Thermostability. Chemical Reviews 2019 119 (3), 1626-1665. https://doi.org/10.1021/acs.chemrev.8b00290