PGxVision (PharmacoGenomic Vision & Interpretation) helps identify and visualize RNA-based cancer biomarkers for drug response. This package is intended to be used in conjunction with the Roche-PharmacoGx pipeline. PGxVision is intended to guide cancer treatment decisions in molecular tumour boards.
R version 4.1.1 (2021-08-10)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19043)To download the package:
require("devtools")
devtools::install_github("EvgeniyaGorobets/PGxVision", build_vignettes=TRUE)
library("PGxVision")To run the shinyApp:
runPGxVision()ls("package:PGxVision")
data(package = "PGxVision")
browseVignettes("PGxVision")PGxVision contains nine functions and four sample data sets. Of the
nine functions, one is to run the Shiny app for this package, three are
plotting functions for visualizing drug sensitivity data, and five are
gene set analysis functions that help provide biological context for
genes of interest.
An overview of the package is illustrated below.
The buildManhattanPlot function visualizes gene response to a certain drugs across the entire human genome. The x-axis represents the full human genome and is labeled by chromosome, and the y-axis can either plot the drug sensitivity or the p-value/fdr/significance of the drug sensitivity statistic. As is the standard for Manhattan plots, data points are colored according to what chromosome they belong to.
The buildVolcanoPlot function lets you simultaneously visualize the drug sensitivity and the p-value/significance of the sensitivity statistic. As is standard for volcano plots, points that are statistically insignificant are grayed out.
The buildWaterfallPlot function plots drug sensitivity (or other similar metrics) across a range of tumours or a range of drugs. The waterfall plot will order results according to the y-axis, so users can quickly identify the most and least sensitive tumour-drug combinations.
You can learn more about how to use these plotting functions in the Plotting Biomarkers section of the Visualizing and Interpreting Biomarkers vignette.
The gene set analysis in this package is broken down into four steps:
- getGeneSets: Given an ENSEMBL gene ID, find all gene sets that contain that gene and return information about those gene sets. Users can examine different types of gene sets, such as those based on biological pathways, cellular components, molecular function, etc.
- expandGeneSets: Using the gene set IDs retrieved in getGeneSets, get all other genes in those gene sets.
- computeGeneSetSimilarity: Using the fully expanded gene sets from expandGeneSets, compute the overlap between the gene sets. All gene sets will at least contain the initial query gene, so pairwise overlap will be non-zero. Currently, the package only supports similarity metrics based on the proportion of intersecting genes to total genes.
- buildNetworkPlot: Plot the gene sets that were retrieved in getGeneSets, using their similarity scores from computeGeneSetSimilarity as edge weights. Gene sets that have high overlap will be in closer proximity and will have thicker and darker edges between them.
In addition to these four functions, the geneSetAnalysis function is provided to run the first three steps of the above pipeline. You can learn more about how to use these gene set analysis functions in the Gene Set Analysis section of the Visualizing and Interpreting Biomarkers vignette.
The runPGxVision lets you perform the functions described above by interacting with UI elements. The dashboard layout of the app lets you see multiple plots side by side, to facilitate analysis. Additionally, the Shiny app lets you interact with the plots (hovering over and selecting points), making it easier to extract useful information.
The screenshots below illustrate what the main features and layout of the app. It is divided into two tabs: “Biomarkers”, which lets you look at gene-level cell-line drug sensitivity data, and “Treatment Response”, which lets you examine and compare treatment responses of different tumours and compounds.
The BRCA.PDXE.paxlitaxel.response data set is a data.frame with
treatment vs. control angle data from BRCA PDXs (patient-derived
xenographs). It is used in the buildWaterfallPlot example and was
retrieved from the Xeva package.
The Biomarkers data set is a data.frame with drug
sensitivity/response data in various experiments involving cancer cell
lines (an experiment is a unique combination of a drug, tissue, and
molecular data type). This data set also contains the p-values for the
drug sensitivity statistics. It is used in the buildManhattanPlot and
buildVolcanoPlot examples and was retrieved from PharmacoDb.
The GRCh38.p13.Assembly data set is a data.frame containing basic
information about the GRCh38.p13 genome assembly. Most notably, it
includes the names and lengths of all the chromosomes. It is used in the
buildManhattanPlot function and was retrieved from Gencode.
The TestGeneSets data set is a data.frame containing 10 different
gene sets and all their constituent genes. This data set represents all
the gene sets in the “GO::CC” (GO cellular compartment) category that
gene ENSG00000012124 is a part of. It is used for testing and was
retrieved from MSigDb.
The author of the package is Evgeniya Gorobets.
data.table is used to transform data.frames into data.tables in
some plotting and gene set analysis functions (buildVolcanoPlot,
buildManhattanPlot, queryGene, expandGeneSets,
computeGeneSetSimilarity). The data.table vignettes were used to
create cleaner syntax and optimize table manipulation.
ggplot2 is used to create non-network plots (buildVolcanoPlot,
buildManhattanPlot, buildWaterfallPlot), and viwNetwork is used to
create network plots (buildNetworkPlot). ggprism is used to enhance
the axes on the Manhattan plot (buildManhattanPlot). viridis is used
to enhance the colors on the network plot (buildNetworkPlot).
checkmate is used to succinctly check user input in all functions.
msigdbr is used to query the MSigDb in gene set analysis functions
(queryGene, expandGeneSets).
shiny, shinydashboard, and shinybusy are used to create the Shiny
app and generate UI elements in the app. plotly is used to add
interactivity to ggplots that are rendered in the Shiny app.
visNetwork is used to add interactivity to visNetwork graphs that are
generated in the Shiny app.
magrittr is used to pipe between functions.
These references describe any and all packages used in PGxVision.
Almende B.V., Thieurmel, B., & Robert, T. (2021). visNetwork: Network Visualization using ‘vis.js’ Library. R package version 2.1.0. https://CRAN.R-project.org/package=visNetwork
Bache, S. M., and Wickham, H. (2020). magrittr: A Forward-Pipe Operator for R. R package version 2.0.1. https://CRAN.R-project.org/package=magrittr
Chang, W. and Borges Ribeiro, B. (2021). shinydashboard: Create Dashboards with ‘Shiny’. R package version 0.7.2. https://CRAN.R-project.org/package=shinydashboard
Chang, W., Cheng, J., Allaire, J., Sievert, C., Schloerke, B., Xie, Y., Allen, J., McPherson, J., Dipert, A., and Borges, B. (2021). shiny: Web Application Framework for R. R package version 1.7.1. https://CRAN.R-project.org/package=shiny
Dawson, C. (2021). ggprism: A ‘ggplot2’ Extension Inspired by ‘GraphPad Prism’. R package version 1.0.3. https://CRAN.R-project.org/package=ggprism
Dolgalev, I. (2021). msigdbr: MSigDB Gene Sets for Multiple Organisms in a Tidy Data Format. R package version 7.4.1. https://CRAN.R-project.org/package=msigdbr
Dowle, M., and Srinivasan, A. (2021). data.table: Extension of
data.frame. R package version 1.14.2.
https://CRAN.R-project.org/package=data.table.
Garnier, S., Ross, N., Rudis, R., Camargo, A. P., Sciaini, M., and Scherer, C. (2021). Rvision - Colorblind-Friendly Color Maps for R. R package version 0.6.2.
Lang, M. (2017). “checkmate: Fast Argument Checks for Defensive R Programming.” The R Journal 9(1), 437-445. https://journal.r-project.org/archive/2017/RJ-2017-028/index.html.
Meyer, F. and Perrier, V. (2020). shinybusy: Busy Indicator for ‘Shiny’ Applications. R package version 0.2.2. https://CRAN.R-project.org/package=shinybusy
R Core Team. (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.
Sievert, C. (2020). Interactive Web-Based Data Visualization with R, plotly, and shiny. Chapman and Hall/CRC.
Wickham, H. (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. https://ggplot2.tidyverse.org.
These references indicate the source for all datasets available in PGxVision.
Feizi, N., Nair, S. K., Smirnov, P., Beri, G., Eeles, C., Esfahani, P. N., … Haibe-Kains, B. (2021). PharmacoDB 2.0: Improving scalability and transparency of in vitro pharmacogenomics analysis. bioRxiv. doi:10.1101/2021.09.21.461211
Frankish, A., Diekhans, M., Ferreira, A. M., Johnson, R., Jungreis, I. Loveland, J., Mudge, J. M., Sisu, C., Wright, J., Armstrong, J., Barnes, I., Berry, A., Bignell, A., Carbonell Sala, S., Chrast, J., Cunningham, F., Di Domenico, T., Donaldson, S., Fiddes, I. T., García Girón, C., … Flicek, P. (2019). GENCODE reference annotation for the human and mouse genomes. Nucleic acids research, 47(D1), D766–D773. https://doi.org/10.1093/nar/gky955
Mer A, Haibe-Kains B (2021). Xeva: Analysis of patient-derived xenograft (PDX) data. R package version 1.10.0.
Subramanian, A., Tamayo, P., Mootha, V. K., Mukherjee, S., Ebert, B. L., Gillette, M. A., … Mesirov, J. P. (2005). Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences, 102(43), 15545–15550. doi:10.1073/pnas.0506580102
These references indicate the source for any code in PGxVision that was taken directly from a website (not including documentation & vignettes for packages).
jdlong. (2018). How to stack two images horizontally in R Markdown. RStudio Community. https://community.rstudio.com/t/how-to-stack-two-images-horizontally-in-r-markdown/18941
Perry & YakovL. (2017). Stop vis.js physics after nodes load but allow drag-able nodes. StackOverflow. https://stackoverflow.com/questions/32403578/stop-vis-js-physics-after-nodes-load-but-allow-drag-able-nodes
This package was developed as part of an assessment for 2021 BCB410H: Applied Bioinformatics, University of Toronto, Toronto, CANADA.