This package provides a simple R interface for Gene Set Enrichment Analysis (GSEA) using the INDRA Biomedical Discovery Engine. The Biomedical Discovery Engine is built on INDRA CoGEx, a graph database integrating causal relations, ontological relations, properties, and data, assembled at scale automatically from the scientific literature and structured sources.
The pre-release version of the package can be pulled from GitHub using the remotes package:
# install.packages("remotes")
remotes::install_github("labsyspharm/r-indra-gsea")Query genes can be either HGNC gene symbols or IDs. The package contains an example set of gene IDs related to COVID-19 infection.
library(indra.gsea)
data("covid_related_genes")
head(covid_related_genes)
#> [1] "613" "1116" "1119" "1697" "7067" "2537"
res <- indra_discrete_gsea(covid_related_genes)The result is a list with the following elements:
names(res)
#> [1] "go" "indra-downstream" "indra-upstream" "phenotype"
#> [5] "reactome" "wikipathways"For example, all entities that are predicted to affect the query genes
upstream are stored in the indra-upstream element:
head(res[["indra-upstream"]], n = 10)| curie | mlp | mlq | name | p | q |
|---|---|---|---|---|---|
| hgnc:5438 | 37.27190 | 32.58177 | IFNG | 5.3e-38 | 2.6e-33 |
| mesh:D007239 | 33.69205 | 29.30295 | Infections | 2e-34 | 5e-30 |
| fplx:Interferon | 31.71050 | 27.49749 | Interferon | 1.9e-32 | 3.2e-28 |
| fplx:IFNA | 29.18266 | 25.09459 | IFNA | 6.6e-30 | 8e-26 |
| hgnc:11892 | 29.05399 | 25.06283 | TNF | 8.8e-30 | 8.7e-26 |
| chebi:16412 | 27.98291 | 24.07094 | lipopolysaccharide | 1e-28 | 8.5e-25 |
| hgnc:5434 | 25.78337 | 21.93834 | IFNB1 | 1.6e-26 | 1.2e-22 |
| hgnc:6018 | 23.76897 | 19.98193 | IL6 | 1.7e-24 | 1e-20 |
| fplx:Protease | 23.22356 | 19.48767 | Protease | 6e-24 | 3.3e-20 |
| hgnc:11362 | 19.87229 | 16.18216 | STAT1 | 1.3e-20 | 6.6e-17 |
These results are acquired by running overrepresentation analysis using Fisher’s exact test and correcting for multiple comparisons using the Benjamini-Hochberg (FDR) method with α = 0.05 on genes causally upstream by one step from all entities in the INDRA database. Analysis was performed using all human genes as the background set.
INDRA Biomedical Discovery Engine was developed by the Gyori Lab at Northeastern University.
INDRA CoGEx (Context Graph Extension) is an automatically assembled biomedical knowledge graph which integrates causal mechanisms from INDRA with non-causal contextual relations including properties, ontologies, and data.
INDRA CoGEx is funded by grant HR00112220036 under the DARPA ASKEM / ARPA-H BDF programs.