Skip to contents

r4msigdb

MSigDB is a widely used gene set database in bio-research. However, navigating and querying pathways of interest on its website can be challenging. This R package is designed to facilitate more convenient and efficient querying of pathways either based on specific genes or using regular expression patterns to match pathway names.

Installation

You can install the development version of r4msigdb like so:

devtools::install('snowGlint/r4msigdb')

Querying Pathways

To retrieve pathways related to a specific topic (e.g., cell programmed death):

library(r4msigdb)
query(species = 'Hs', pathway = 'OPTOSIS') |> head()
#> Key: <collection_name, standard_name>
#>    collection_name                              standard_name
#>             <char>                                     <char>
#> 1:          C2:CGP                           ALCALA_APOPTOSIS
#> 2:          C2:CGP           BROCKE_APOPTOSIS_REVERSED_BY_IL6
#> 3:          C2:CGP       CONCANNON_APOPTOSIS_BY_EPOXOMICIN_DN
#> 4:          C2:CGP       CONCANNON_APOPTOSIS_BY_EPOXOMICIN_UP
#> 5:          C2:CGP DEBIASI_APOPTOSIS_BY_REOVIRUS_INFECTION_DN
#> 6:          C2:CGP DEBIASI_APOPTOSIS_BY_REOVIRUS_INFECTION_UP
#>                                          symbol
#>                                          <list>
#> 1:         HCCS,MATK,FAS,CYFIP2,ELOVL1,PFKP,...
#> 2:      DPM1,RALA,PHTF2,ADIPOR2,CD44,SH2D2A,...
#> 3: ICA1,ETV1,TRAPPC6A,DNASE1L1,TMSB10,HDAC9,...
#> 4:          TAC1,IFRD1,TSPAN9,GCLM,FAS,CD44,...
#> 5:      LASP1,BLTP2,METTL13,CD9,NISCH,BRCA1,...
#> 6:           AK2,CDC27,ACSM3,ZFX,TAC1,IFRD1,...

This will return a list of pathways associated with OPTOSIS and the genes involved in each pathway.

If you want to find pathways related to specific genes (e.g., PTPRC and TP53):

query(species = 'Hs', symbols = c('PTPRC', 'TP53')) |> head()
#> Key: <collection_name, standard_name>
#>    collection_name                                   standard_name
#>             <char>                                          <char>
#> 1:              C1                                        chr17p13
#> 2:              C1                                         chr1q31
#> 3:          C2:CGP             ABRAHAM_ALPC_VS_MULTIPLE_MYELOMA_DN
#> 4:          C2:CGP ACOSTA_PROLIFERATION_INDEPENDENT_MYC_TARGETS_DN
#> 5:          C2:CGP      BENITEZ_GBM_PROTEASOME_INHIBITION_RESPONSE
#> 6:          C2:CGP                       BENPORATH_MYC_MAX_TARGETS
#>                                        symbol
#>                                        <list>
#> 1: CAMKK1,DVL2,DHX33,PAFAH1B1,GAS7,RABEP1,...
#> 2:        GLRX2,TPR,PTGS2,RGS2,RO60,UCHL5,...
#> 3:      TCF3,XRCC5,PSEN1,CCNB1,IL6ST,CDK4,...
#> 4:     LASP1,ACSM3,GDE1,CAPN1,CHPF2,KCNH2,...
#> 5:     DVL2,MXD1,ASNS,MLF2,GPATCH2L,FBXO7,...
#> 6:        DPM1,GCLC,M6PR,RECQL,GCFC2,PDK2,...

This will provide pathways where the specified genes are involved.

Custom Query

Advanced users familiar with data.table can perform custom queries. For example, to retrieve all pathways collection_name == 'H':

query(species = 'Hs')[.('H')] |> head()
#> Key: <collection_name, standard_name>
#>    collection_name                standard_name
#>             <char>                       <char>
#> 1:               H        HALLMARK_ADIPOGENESIS
#> 2:               H HALLMARK_ALLOGRAFT_REJECTION
#> 3:               H   HALLMARK_ANDROGEN_RESPONSE
#> 4:               H        HALLMARK_ANGIOGENESIS
#> 5:               H     HALLMARK_APICAL_JUNCTION
#> 6:               H      HALLMARK_APICAL_SURFACE
#>                                           symbol
#>                                           <list>
#> 1:  AK2,NDUFAB1,ADIPOR2,UQCRC1,PHLDB1,RETSAT,...
#> 2:            STAB1,BRCA1,WAS,FAS,CAPG,HDAC9,...
#> 3:        PGM3,PIAS1,RRP12,APPBP2,GNAI3,IDI1,...
#> 4:           NRP1,JAG1,TIMP1,VTN,VEGFA,ITGAV,...
#> 5:          CD99,SKAP2,ITGA3,PKD1,CLDN11,VCL,...
#> 6: ADIPOR2,BRCA1,DCBLD2,CROCC,IL2RB,ATP6V0A4,...

You can also use .unlist = TRUE to unlist the symbols column in any of the above query methods.

query(species = 'Hs', .unlist = TRUE) |> head()
#> Key: <collection_name, standard_name>
#>    collection_name standard_name  symbol
#>             <char>        <char>  <char>
#> 1:              C1            MT  MT-CO2
#> 2:              C1            MT  MT-ND2
#> 3:              C1            MT  MT-CO1
#> 4:              C1            MT  MT-ND3
#> 5:              C1            MT  MT-ND4
#> 6:              C1            MT MT-ATP6

GSEA 

library(fgsea)
library(data.table)
palette <- c("#440154FF", "#31688EFF", "#26828EFF", "#6DCD59FF", "#FDE725FF")

data(exampleRanks)
pathway <-  query(species = 'Hs', pathway = 'OPTOSIS', .unlist = TRUE)
pathway <- pathway[, .(standard_name, symbol)]
# substitute names with random symbols
set.seed(2024)
names(exampleRanks) <- query(species = 'Hs', .unlist = TRUE)[, sample(unique(symbol), length(exampleRanks))]

gseaR <- clusterProfiler::GSEA(rev(exampleRanks),
  TERM2GENE = pathway, pvalueCutoff = 1, by = "fgsea", eps = 0
)
sortedgsea <- as.data.table(gseaR@result)[order(pvalue)]

enrichplot::gseaplot2(gseaR, sortedgsea[["ID"]][1:5],
  base_size = 10,
  color = palette,
  rel_heights = c(1.5, 0.3, 0.5),
  pvalue_table = FALSE
)