Question

runSPIA() in graphite package

0

Entering edit mode

array chip ▴ 420

@array-chip-4136

Last seen 12 weeks ago

United States

Hi all, I am using runSPIA() from graphite package to analyze a gene list against biocarta pathway database. > library(graphite) > prepareSPIA(biocarta, "biocartaEx") > obj<-runSPIA(de=siggenes, all=allgenes, "biocartaEx") where "siggenes" are a list of 1332 significant genes selected from "allgenes" (a list of 17138 genes). The running process verbose indicated a total of 178 pathways analyzed. One of the pathways I am particularly interested is "estrogen responsive protein efp controls cell cycle and breast tumors growth": > p <- biocarta[["estrogen responsive protein efp controls cell cycle and breast tumors growth"]] > nodes(p) [1] "CDKs"???????????? "Cyclin B1/2"????? "EntrezGene:2099"? "EntrezGene:2810"? "EntrezGene:57154" "EntrezGene:7157"? "EntrezGene:7706"?? "ubiquitin" And my siggenes contains 2 of genes involved in this pathway: > siggenes[c('7706','2099')] ???? 7706????? 2099 ?4.347012 -3.792425 So I would assume that runSPIA will examine this pathway during the? calculation, but surprisingly I didn't see this particular way being examined. Here is the section of runSPIA() verbose output that started with "e": Done pathway 42 : e2f1 destruction pathway.. Done pathway 43 : effects of calcineurin in kera.. Done pathway 44 : egf signaling pathway.. Done pathway 45 : endocytotic role of ndk phosph.. Done pathway 46 : epo signaling pathway.. Done pathway 47 : erk and pi-3 kinase are necess.. Done pathway 48 : erk1/erk2 mapk signaling pathw.. Done pathway 49 : eukaryotic protein translation.. Done pathway 50 : extrinsic prothrombin activati.. Can anyone tell me why runSPIA() missed this pathway? Attached are siggenes and allgenes for you to try.? > siggenes<-as.matrix(read.table("siggenes.txt",row.names=1))[,1] > allgenes<-as.matrix(read.table("H:\\test\\allgenes.txt", row.names=NULL, header=T, colClasses='character'))[,1] Many thanks John -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: allgenes.txt URL: <https: stat.ethz.ch="" pipermail="" bioconductor="" attachments="" 20130208="" 1a59d276="" attachment.txt=""> -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: siggenes.txt URL: <https: stat.ethz.ch="" pipermail="" bioconductor="" attachments="" 20130208="" 1a59d276="" attachment-0001.txt="">

Pathways Breast siggenes PROcess cycle graphite Pathways Breast siggenes PROcess cycle • 1.9k views

ADD COMMENT • link updated 11.2 years ago by Enrica ▴ 20 • written 11.2 years ago by array chip ▴ 420

score 0 · Answer 1 · 2013-02-11

Dear Enrica, Thanks very much for checking this out! graphite is a great package for pathway analysis with its ability to analyze so many different pathway databases! One more question, does runSPIA() run against these pathways on the fly (i.e. access these databases from their server in real time), or against a pre-downloaded database? Do you have a timeline when the new release graphite package will become available? Thanks again for your great work. John ________________________________ From: Enrica <enrica.calura@gmail.com> Cc: Bioconductor mailing list <bioconductor@stat.math.ethz.ch>; "gabriele.sales@unipd.it" <gabriele.sales@unipd.it>; "chiara.romualdi@unipd.it" <chiara.romualdi@unipd.it> Sent: Monday, February 11, 2013 3:24 AM Subject: Re: runSPIA() in graphite package Dear John, the procedure you applied to analyse your data is right. However, the runSpia function filters out those pathways that have no edges or less than 5 nodes with valid edges. Your pathway, after the conversion to Entrez Gene, do not satisfy neither the conditions, having no edges. > pe<-convertIdentifiers(biocarta[["estrogen responsive protein efp controls cell cycle and breast tumors growth"]], "entrez") > pe "estrogen responsive protein efp controls cell cycle and breast tumors growth" pathway from BioCarta Number of nodes = 5 Number of edges = 0 Type of identifiers = Entrez Gene Retrieved on = 2011-05-12 We equipped our runSpia() function with the checks described above in order to protect the user from unreliable results. On a separate note, we have also re-checked how that specific pathway was converted. The original BioPax2 contained some edges, but their endpoints were unfortunately nodes which could not be associated with any entrez gene. Our software, thus, dropped them from the pathway. We are working on a new graphite release based on annotation released more recently. Those includes more edges; as a result that pathway is no longer empty. Enrica Calura Hi all, I am using runSPIA() from graphite package to analyze a gene list against biocarta pathway database. > > > >> library(graphite) > >> prepareSPIA(biocarta, "biocartaEx") >> obj<-runSPIA(de=siggenes, all=allgenes, "biocartaEx") > > > >where "siggenes" are a list of 1332 significant genes selected from "allgenes" (a list of 17138 genes). The running process verbose indicated a total of 178 pathways analyzed. > > > >One of the pathways I am particularly interested is "estrogen responsive protein efp controls cell cycle and breast tumors growth": > > >> p <- biocarta[["estrogen responsive protein efp controls cell cycle and breast tumors growth"]] >> nodes(p) >[1] "CDKs" "Cyclin B1/2" "EntrezGene:2099" "EntrezGene:2810" "EntrezGene:57154" "EntrezGene:7157" "EntrezGene:7706" "ubiquitin" > > >And my siggenes contains 2 of genes involved in this pathway: > > >> siggenes[c('7706','2099')] > 7706 2099 > 4.347012 -3.792425 > > >So I would assume that runSPIA will examine this pathway during the calculation, but surprisingly I didn't see this particular way being examined. Here is the section of runSPIA() verbose output that started with "e": > > >Done pathway 42 : e2f1 destruction pathway.. >Done pathway 43 : effects of calcineurin in kera.. >Done pathway 44 : egf signaling pathway.. >Done pathway 45 : endocytotic role of ndk phosph.. >Done pathway 46 : epo signaling pathway.. >Done pathway 47 : erk and pi-3 kinase are necess.. >Done pathway 48 : erk1/erk2 mapk signaling pathw.. >Done pathway 49 : eukaryotic protein translation.. >Done pathway 50 : extrinsic prothrombin activati.. > > > >Can anyone tell me why runSPIA() missed this pathway? Attached are siggenes and allgenes for you to try. > > >> siggenes<-as.matrix(read.table("siggenes.txt",row.names=1))[,1] > >> allgenes<-as.matrix(read.table("H:\\test\\allgenes.txt", row.names=NULL, header=T, colClasses='character'))[,1] > > >Many thanks > > >John > > [[alternative HTML version deleted]]

score 0 · Answer 2 · 2013-02-11

Dear John, the procedure you applied to analyse your data is right. However, the runSpia function filters out those pathways that have no edges or less than 5 nodes with valid edges. Your pathway, after the conversion to Entrez Gene, do not satisfy neither the conditions, having no edges. > pe<-convertIdentifiers(biocarta[["estrogen responsive protein efp controls cell cycle and breast tumors growth"]], "entrez") > pe "estrogen responsive protein efp controls cell cycle and breast tumors growth" pathway from BioCarta Number of nodes = 5 Number of edges = 0 Type of identifiers = Entrez Gene Retrieved on = 2011-05-12 We equipped our runSpia() function with the checks described above in order to protect the user from unreliable results. On a separate note, we have also re-checked how that specific pathway was converted. The original BioPax2 contained some edges, but their endpoints were unfortunately nodes which could not be associated with any entrez gene. Our software, thus, dropped them from the pathway. We are working on a new graphite release based on annotation released more recently. Those includes more edges; as a result that pathway is no longer empty. Enrica Calura 2013/2/8 array chip <arrayprofile@yahoo.com> > Hi all, I am using runSPIA() from graphite package to analyze a gene list > against biocarta pathway database. > > > library(graphite) > > prepareSPIA(biocarta, "biocartaEx") > > obj<-runSPIA(de=siggenes, all=allgenes, "biocartaEx") > > where "siggenes" are a list of 1332 significant genes selected from > "allgenes" (a list of 17138 genes). The running process verbose indicated a > total of 178 pathways analyzed. > > One of the pathways I am particularly interested is "estrogen responsive > protein efp controls cell cycle and breast tumors growth": > > > p <- biocarta[["estrogen responsive protein efp controls cell cycle and > breast tumors growth"]] > > nodes(p) > [1] "CDKs" "Cyclin B1/2" "EntrezGene:2099" > "EntrezGene:2810" "EntrezGene:57154" "EntrezGene:7157" > "EntrezGene:7706" "ubiquitin" > > And my siggenes contains 2 of genes involved in this pathway: > > > siggenes[c('7706','2099')] > 7706 2099 > 4.347012 -3.792425 > > So I would assume that runSPIA will examine this pathway during the > calculation, but surprisingly I didn't see this particular way being > examined. Here is the section of runSPIA() verbose output that started with > "e": > > Done pathway 42 : e2f1 destruction pathway.. > Done pathway 43 : effects of calcineurin in kera.. > Done pathway 44 : egf signaling pathway.. > Done pathway 45 : endocytotic role of ndk phosph.. > Done pathway 46 : epo signaling pathway.. > Done pathway 47 : erk and pi-3 kinase are necess.. > Done pathway 48 : erk1/erk2 mapk signaling pathw.. > Done pathway 49 : eukaryotic protein translation.. > Done pathway 50 : extrinsic prothrombin activati.. > > Can anyone tell me why runSPIA() missed this pathway? Attached are > siggenes and allgenes for you to try. > > > siggenes<-as.matrix(read.table("siggenes.txt",row.names=1))[,1] > > allgenes<-as.matrix(read.table("H:\\test\\allgenes.txt", row.names=NULL, > header=T, colClasses='character'))[,1] > > Many thanks > > John > > [[alternative HTML version deleted]]