README.md

In travis-m-blimkie/networker: Reproducible PPI Network Creation and Visualization in R

networker: Reproducible PPI Network Creation and Visualization in R

A suite of functions to make it simple to construct PPI networks inside of R, with an emphasis on usability and reproducibility.

Installation

You can install networker with the following (assuming you have devtools already installed):

devtools::install_github("https://github.com/travis-m-blimkie/networker")

If you run into installation issues relating to the package biomaRt or ReactomePA (which are dependencies of networker), you can install them with the following:

if (!require("BiocManager")) { # Optionally install BiocManager if needed
  install.packages("BiocManager")
}
BiocManager::install("biomaRt", "ReactomePA")

Example

Start off by loading the package, reading in some genes, and building a network using the "min_steiner" option, which works well for most small- to mid-sized networks.

> library(networker)
# Thanks for using networker! If you encounter any bugs or problems, please 
# submit an issue at the Github page: 
# https://github.com/travis-m-blimkie/networker/issues

> ex_genes <- read_csv("ex_de_genes.csv")
> glimpse(ex_genes)
# Rows: 158
# Columns: 3
# $ ensembl_gene_id <chr> "ENSG00000160654", "ENSG00000018280", "ENSG00000179...
# $ hgnc_symbol     <chr> "CD3G", "SLC11A1", "CIITA", "EIF4E3", "RNASE6", "AL...
# $ log2FoldChange  <dbl> 0.07662557, 1.02935351, -2.59604031, -0.47667505, 2...

> ex_network <- build_network(
    df    = ex_genes,
    col   = "ensembl_gene_id",
    order = "min_steiner",
    hub_measure = "degree"
  )
# ==> INFO: Found 3 duplicate IDs in the input column, which have been removed:
# ENSG00000172724, ENSG00000169397, ENSG00000169385
# 
# Finding interactions...
# Creating network...
# Performing 'Steiner' minimum network trimming...
# Done.

Notice that duplicate IDs are automatically removed from the input, with a message to the user. Now we can visualize the network with plot_network, using the automatically added "gene_name" column to provide the labels:

> plot_network(
    network      = ex_net,
    fill_column  = log2FoldChange,
    fill_type    = "fold_change",
    layout       = "nicely",
    label        = TRUE,
    label_column = gene_name,
    label_filter = 2,
    label_size   = 4,
    legend       = TRUE
  )
# Warning message:
# Removed 166 rows containing missing values (geom_text_repel).

We can then use the enrich_network to test for significantly enriched Reactome pathways (implemented with ReactomePA):

> (sig_pathways <- enrich_network(network = ex_net, filter = 0.05))
# # A tibble: 211 × 5
#    id            description                                          p_value   p_adjust gene_id
#    <chr>         <chr>                                                  <dbl>      <dbl> <chr>  
#  1 R-HSA-2029480 Fcgamma receptor (FCGR) dependent phagocytosis       2.15e-11    1.01e-8 695/74…
#  2 R-HSA-449147  Signaling by Interleukins                            2.22e-11    1.01e-8 3383/8…
#  3 R-HSA-2424491 DAP12 signaling                                      2.72e-10    8.16e-8 695/58…
#  4 R-HSA-1433557 Signaling by SCF-KIT                                 4.33e-10    8.16e-8 5879/5…
#  5 R-HSA-877300  Interferon gamma signaling                           6.24e-10    8.16e-8 3383/5…
#  6 R-HSA-9664407 Parasite infection                                   7.18e-10    8.16e-8 695/74…
#  7 R-HSA-9664417 Leishmania phagocytosis                              7.18e-10    8.16e-8 695/74…
#  8 R-HSA-9664422 FCGR3A-mediated phagocytosis                         7.18e-10    8.16e-8 695/74…
#  9 R-HSA-2029482 Regulation of actin dynamics for phagocytic cup ...  1.05e- 9    1.06e-7 695/74…
# 10 R-HSA-388841  Costimulation by the CD28 family                     4.13e- 9    3.75e-7 5879/5…
# # … with 201 more rows

Finally, we can pick one of the pathways output above, and pull out a module/subnetwork based on the genes belonging to that pathway, which are highlighted when plotting:

> ex_net_module <- extract_subnetwork(
    network       = ex_net,
    enrich_result = sig_pathways,
    pathway_name  = "Interferon gamma signaling"
  )
# Checking inputs...
# Pulling genes for given pathway...found 13 genes.
# Calculating subgraphs from specified nodes...
# Determining shortest paths between nodes...
# Done, new subnetwork contains 20 nodes.

> plot_network(
    network      = ex_net_module,
    fill_column  = log2FoldChange,
    fill_type    = "fold_change",
    legend       = TRUE,
    layout       = "nicely",
    label        = TRUE,
    label_column = gene_name,
    label_filter = 0,
    subnet       = TRUE
  )
# Detected this is a sub-network generated by `extract_subnetwork()`.
# Highlighted node labels indicate genes from the extracted pathway.

Versioning

This package makes use of SemVer.

Authors

Travis Blimkie is the originator and principal contributor. You can check the list of all contributors here.

License

This project is written under the GPLv3 license, available here.

travis-m-blimkie/networker documentation built on June 3, 2023, 10:17 p.m.