cordial

A Pleasant Tonic For Parallel Correlation Analysis In R

Computes pairwise Pearson's correlations of a dataset, or specified targets simultaneously in parallel. Additionally, fits a linear model. Conveniently includes the ability to filter the input dataset and select a subset of columns to compute correlations. Outputs Pearson's product moment correlation coefficients, p-values, adjusted p-values, linear model slope and observation counts in long-format.

Installation

RStudio

1. Download the Package Archive File (cordial_x.x.x.tar.gz) of the latest release.
   • Not the Source code
2. In RStudio, click: Tools menu
3. Select: Install Packages...
4. In the Install from list box, select: Package Archive File (.zip; .tar.gz)
5. Click: Browse to select the downloaded cordial Package Archive File
6. Select: Install

R console

One-liner

Run:

devtools::install_url(
 "https://github.com/CutillasLab/cordial/releases/download/v0.1.10/cordial_0.1.10.tar.gz"
)

• You may first need to install devtools: install.packages("devtools")

Manual

1. Download the Package Archive File (cordial_x.x.x.tar.gz) of the latest release.
   • Not the Source code
2. Run: devtools::install_local(path = "C:/path/to/cordial_x.x.x.tar.gz")
   • Replace the string argument to path with the actual location
   • You may first need to install devtools: install.packages("devtools")

Notes

• The source files in cordial/data/ does NOT contain the crispr_DT data, as it exceeds the GitHub file size limit.
• The .tar.gz in the assets of the latest release DOES contain the data.
• Due to the excessive size of included datasets, Git Large File Storage (LFS) has been implemented. This causes known issues when installing packages directly from GitHub (bad restore file magic number (file may be corrupted) -- no data loaded); therefore, installations must use the Package Archive File (cordial_x.x.x.tar.gz) from the latest release.
• If installation error occurs stating that certain required packages are missing, try manually installing the missing packages listed prior to reattempting to install cordial, e.g.:

devtools::install_cran(c("ggplot2", "ggrepel", "magrittr", "tidyr", "purrr", "future", "furrr", "collapse", "data.table"))

• If error persists, it may also be required to install additional build tools specific to your operating system (OS) and R version:
  • Windows: https://cran.r-project.org/bin/windows/Rtools/
  • macOS: https://cran.r-project.org/bin/macosx/tools
  • Linux: https://cran.r-project.org/bin/linux/

Quickstart

Load

• To load cordial:
  • library(cordial)
• After loading the package, explore the help documentation by entering into the console the name of the package or function prefixed with a question mark, e.g.:
  • ?cordial
  • ?cor_map

Functions

cor_map()

To get pairwise correlations for all target (column) permutations.

• Parameters:
  • dataset: A wide-format data.table.
  • select_cols: A vector of column names (character), or indices (numeric) to select; must omit non-numeric columns.
  • filter_rows: A named list. Values specify which rows to subset. Names correspond to column names in dataset, or metadata if supplied.
  • metadata: An optional data.table with values corresponding to rows matching in dataset.
  • self: A character string. "yes" includes self-correlations; "no" omits.
  • method: A character string. Correction method for p-value adjustment, passed to stats::p.adjust().

cor_target_map()

To get all correlations for a selection of key targets.

• Supply the key targets as a vector (e.g., c("COLUMN1", "COLUMN2")) to the additional target parameter.
• Other parameters are shared with cor_map().

cor_target()

To get all correlations for a single key target.

• Supply a character string of the key target (e.g., "COLUMN1") to the additional target parameter.
• Other parameters are shared with cor_map().

Notes

Correlations for all columns will be calculated, unless a subset is supplied to select_cols, i.e.:

• corr(X, Y) = corr(select_cols, select_cols)

The additional target parameter (cor_target_map(), cor_target()) allows further subsetting of the pairwise correlations to make, i.e.:

• corr(X, Y) = corr(target, select_cols)

References

Badshah II, Cutillas PR. Systematic identification of biochemical networks in cancer cells by functional pathway inference analysis. Bioinformatics. 2023 Jan 1;39(1):btac769. PMID: 36448701; PMCID: PMC9805595. https://doi.org/10.1093/bioinformatics/btac769