R: download all KEGG pathways including KO and Compounds

Question

[dago]

I saw this question has been asked here and there before. However, I could not find a tool that does the job for me.

I want to download all pathways from KEGG including KO and compounds using R. I would imagine creating an R object like:

$Path_1
...KO
...Compounds
$Path_2
...KO
...Compounds
$Path_3
...KO
...Compounds

Any idea how to download the data?

Thank you

[GenoMax]

all pathways from KEGG including KO and compounds using R.

That would violate their AUP if you don't have a license.

[dago]

I did not think about this. I guess I an getting used to have open source tools/db. Thanks

Answer 1

[5heikki]

You can use their API. However, it is not meant for downloading the entire database. For that there is the ftp which is behind a license

[dago]

ah right, that is maybe why I could not find any tool doing that!

Answer 2

[ATpoint]

MSigDB contains the KEGG pathways: https://www.gsea-msigdb.org/gsea/msigdb/collections.jsp Download the gmt file and then load it into R, e.g. with

kegg <- fgsea::gmtPathways("c2.cp.kegg.v7.1.symbols.gmt")

> head(kegg)
$KEGG_GLYCOLYSIS_GLUCONEOGENESIS
 [1] "ACSS2"   "GCK"     "PGK2"    "PGK1"    "PDHB"    "PDHA1"   "PDHA2"   "PGM2"   
 [9] "TPI1"    "ACSS1"   "FBP1"    "ADH1B"   "HK2"     "ADH1C"   "HK1"     "HK3"    
[17] "ADH4"    "PGAM2"   "ADH5"    "PGAM1"   "ADH1A"   "ALDOC"   "ALDH7A1" "LDHAL6B"
[25] "PKLR"    "LDHAL6A" "ENO1"    "PKM"     "PFKP"    "BPGM"    "PCK2"    "PCK1"   
[33] "ALDH1B1" "ALDH2"   "ALDH3A1" "AKR1A1"  "FBP2"    "PFKM"    "PFKL"    "LDHC"   
[41] "GAPDH"   "ENO3"    "ENO2"    "PGAM4"   "ADH7"    "ADH6"    "LDHB"    "ALDH1A3"
[49] "ALDH3B1" "ALDH3B2" "ALDH9A1" "ALDH3A2" "GALM"    "ALDOA"   "DLD"     "DLAT"   
[57] "ALDOB"   "G6PC2"   "LDHA"    "G6PC"    "PGM1"    "GPI"    

$KEGG_CITRATE_CYCLE_TCA_CYCLE
 [1] "IDH3B"    "DLST"     "PCK2"     "CS"       "PDHB"     "PCK1"     "PDHA1"   
 [8] "PDHA2"    "SUCLG2P2" "FH"       "SDHD"     "OGDH"     "SDHB"     "IDH3A"   
[15] "SDHC"     "IDH2"     "IDH1"     "ACO1"     "ACLY"     "MDH2"     "DLD"     
[22] "MDH1"     "DLAT"     "OGDHL"    "PC"       "SDHA"     "SUCLG1"   "SUCLA2"  
[29] "SUCLG2"   "IDH3G"    "ACO2"    

$KEGG_PENTOSE_PHOSPHATE_PATHWAY
 [1] "RPE"     "RPIA"    "PGM2"    "PGLS"    "PRPS2"   "FBP2"    "PFKM"    "PFKL"   
 [9] "TALDO1"  "TKT"     "FBP1"    "TKTL2"   "PGD"     "RBKS"    "ALDOA"   "ALDOC"  
[17] "ALDOB"   "H6PD"    "RPEL1"   "PRPS1L1" "PRPS1"   "DERA"    "G6PD"    "PGM1"   
[25] "TKTL1"   "PFKP"    "GPI"

[dago]

That is actually great! But I am not sure there are compounds here, just name of genes. No?

[andres.firrincieli]

not the best solution because is for single organisms, but genome scale metabolic models (http://bigg.ucsd.edu/data_access) have all the information you need regarding the Gene-Protein-Reaction associations. Once you have the gene id, getting the KO with eggNOG shoudl not be a problem.