ProtConservR

R package for Analyzing and Visualizing Protein Sequence Conservation.

Description

ProtConservR is an R package for analyzing and visualizing protein sequence conservation. It provides functions to (1) Read and process protein sequences from FASTA files, (2) Create multiple sequence alignments from these protein sequence multifasta files using MUSCLE or ClustalW. (3) Calculate position-specific conservation scores based on shannon entropy. (4) Map sequence conservation scores to protein structure data. (5) Create plots and interactive visualizations of conservation data, including line plots of conservation scores across residues, heatmaps of conservation scores on protein structure data (residues), and 3D visualizations of protein structure colored by conservation.

Session Info: R version 4.3.1 (2023-06-16) Platform: aarch64-apple-darwin20 (64-bit) Running under: macOS Sonoma 14.6.1

Installation

You can install the latest version of ProtConservR like so:

install.packages("devtools")
library("devtools")
devtools::install_github("dhanyajagan/ProtConservR", build_vignettes = TRUE)
library("ProtConservR")

To run the Shiny app: runProtConservR()

Overview

The instructions below are for taking a look at the vignettes for this package:

ls("package:ProtConservR")
data(package = "ProtConservR") 
browseVignettes("ProtConservR")

ProtConservR contains 8 functions.

1. readSequences Reads in unaligned protein sequences provided in a multifasta file. Protein sequences should be of a particular protein family, like hemoglobin, insulin,etc. User can use sample data located in inst/extdata/, or they can pull multifasta files of protein families of their choice from InterPro website as input into this function.

2. createAlignment Performs multiple sequence alignment on protein sequences using either MUSCLE or ClustalW algorithm with default settings. Takes processed sequences from readSequences() function as input.

3. calculateConservation This function calculates conservation scores for each position in a multiple sequence alignment using Shannon’s entropy. The scores are then normalized to a 0-1 scale where 1 indicates perfect conservation.

4. plotConservationScores This function creates a plot of conservation scores for each position in a multiple sequence alignment (plotting sequence conservation scores).

5. readStructure This function reads and processes a protein structure from a specified PDB file. Users must provide a file path to a PDB file.

6. mapConservation This function maps sequence conservation scores onto protein structure by aligning the structural residues with the corresponding positions in the multiple sequence alignment. It handles potential gaps and mismatches between the structure and alignment in a very simple way.

7. plotInteractiveConservation Creates an interactive visualization of protein conservation scores using plotly. This function takes as input the output of mapConservation() and generates multiple visualization types including a heatmap view of the conservation scores, a line plot showing conservation across residues, and summary statistics.

8. threeDimStructureConservation Creates an interactive 3D visualization of the pdb protein structure colored by conservation scores. Takes as input the representative pdb file and output from mapConservation and uses r3dmol to create an interactive viewer where conservation scores are mapped to a color gradient from red (low conservation) to white (medium) to blue (high conservation).

Refer to package vignettes for more details. An overview of the package is illustrated below.

Contributions

The author of the package is Dhanya Jagannathan. The author wrote the following 8 functions: readsequences, readStructure, createAlignment, calculateConservation, mapConservation, plotConservationScores, plotInteractiveConservation, and threeDimStructureConservation. The readsequences function makes use of the Biostrings package to read in fasta file input, and takes inspiration from the bio3D package’s read.fasta() function. The createAlignment function makes use of the Biostrings, msa, and muscle packages, and takes inspiration from the bio3D package’s read.fasta() function. The calculateConservation function makes use of Biostrings package, and takes inspiration from the bio3D package’s conserv() and consensus() functions. It also takes inspiration from the fact that shannon entropy was used in bio3D package as well. However, this packages implementation to calculate conservation scores utilized the Biostrings consensusmatrix, which is what makes it different. The plotConservationScores function makes heavy use of ggplot2 package for plotting of the conservation scores from alignment. The readStructure function to read in pdb protein files utilizes the read.pdb() function from the bio3D package. The mapConservation function utilizes the bio3D package functions atom.select and pdbseq in order to map the conservation scores from protein sequence alignment to the pdb structure/residues. The plotInteractiveConservation function makes interactive plots of the output of mapConservation, and utilizes the plotly, stats, and magrittr packages. Lastly, the function threeDimStructureConservation utilizes the r3dmol and grDevices packages to create an interactive 3D protein visualization of sequence conservation overlayed onto protein structure. ChatGPT was used across the functions, to develop the roxygen tags for the functions, and was also used to debug errors caused by incorrectly written functions. Other than this, generative AI was NOT used for core code/ideas/problem solving.

References

• Nicholas Rego and David Koes, 3Dmol.js: molecular visualization with WebGL,, Volume 31, Issue 8, April 2015, Pages 1322–1324, .

• Sievert C (2020). Interactive Web-Based Data Visualization with R, plotly, and shiny. Chapman and Hall/CRC. ISBN 9781138331457, https://plotly-r.com.

• Barry J. Grant, Ana P. C. Rodrigues, Karim M. ElSawy, J. Andrew McCammon, Leo S. D. Caves, Bio3d: an R package for the comparative analysis of protein structures, , Volume 22, Issue 21, November 2006, Pages 2695–2696, .

• Wickham H (2016). Springer-Verlag New York. ISBN 978-3-319-24277-4, https://ggplot2.tidyverse.org.

• Valdar, W.S.J. (2002). Scoring residue conservation. , 48(2), 227-241.

• Strait BJ, Dewey TG. The Shannon information entropy of protein sequences. Biophys J. 1996 Jul;71(1):148-55. doi: 10.1016/S0006-3495(96)79210-X. PMID: 8804598; PMCID: PMC1233466.

• Pagès H, Aboyoun P, Gentleman R, DebRoy S (2024). R package version 2.74.0, .

• Edgar, R.C. (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. 32(5):1792-1797.

• Thompson, J.D., Higgins, D.G., Gibson, T.J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. 22(22):4673-4680.

• Bodenhofer U, Bonatesta E, Horejs-Kainrath C, Hochreiter S (2015). “msa: an R package for multiple sequence alignment.” , 31(24), 3997–3999. doi:10.1093/bioinformatics/btv494

• OpenAI. (2024). ChatGPT (November 5, 2024). Retrieved from https://chat.openai.com

Acknowledgements

This package was developed as part of an assessment for 2024 BCB410H: Applied Bioinformatics course at the University of Toronto, Toronto, CANADA. ProtConservR welcomes issues, enhancement requests, and other contributions. To submit an issue, use theGitHub issues