Dear all, This might not be a specific BioConductor question. However, BioConductor is my major tool and therefore I ask the question here. If not appropriate please ignore. I am analyzing array data from a set of mutants all showing the same phenotype. The idea was of course to find genes similarly expressed in all mutants to get closer to the regulatory mechanism of the phenotype. Surprisingly there was nearly no overlap among the differentially expressed genes. The suggestion was raised that the genes might encode similar functions. I therefore conducted GO analysis (using topGO and other packages). Still very little obvious overlap but how to quantify it? How to determine the degree of similarity between different GO results? All suggestions are welcome. Johannes Hanson Molecular Plant Physiology, Utrecht University

Hi Johannes There are several packages in BioC that you can use to do GO analysis or to mine GO analysis results. You can find most, if not all, of them in the Bioconductor GO task view. (Curiously topGO is not there). http://www.bioconductor.org/packages/release/GO.html I am the maintainer of the 'goProfiles' package which is, in principle, developed to answer the type of question you pose. If you have any questions about its use I'll be pleased to try to help you at answering them. best Alex S?nchez ...................................................................... ......................................... Dr. Alex S?nchez. Associate Professor. Statistics Department. University of Barcelona. Facultat de Biologia UB. Avda Diagonal 645. 08028 Barcelona. Spain asanchez_at_ub.edu Statistics and Bioinformatics Unit Institut de Recerca. Hospital Universitari Vall 'Hebron Passeig Vall d'Hebron 112-119. 08034 Barcelona asanchez_at_ir.vhebron.net ...................................................................... ......................................... > Message: 17 > Date: Wed, 12 Aug 2009 09:10:56 +0200 > From: Johannes Hanson <s.j.hanson at="" uu.nl=""> > Subject: [BioC] How to compare GO results > To: bioconductor at stat.math.ethz.ch > Message-ID: <02C00589-9261-449C-B8A4-D7C8C4986E72 at uu.nl> > Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes > > Dear all, > > This might not be a specific BioConductor question. However, > BioConductor is my major tool and therefore I ask the question here. > If not appropriate please ignore. > > I am analyzing array data from a set of mutants all showing the same > phenotype. The idea was of course to find genes similarly expressed in > all mutants to get closer to the regulatory mechanism of the phenotype. > Surprisingly there was nearly no overlap among the differentially > expressed genes. The suggestion was raised that the genes might encode > similar functions. I therefore conducted GO analysis (using topGO and > other packages). Still very little obvious overlap but how to quantify > it? How to determine the degree of similarity between different GO > results? > > All suggestions are welcome. > > Johannes Hanson > Molecular Plant Physiology, Utrecht University > > >

Hello goProfiles works with 'organism' anotation packages ('org.XXX.XXX.db') There is an Arabidopsis annotation package 'org.At.tair.db' in the next version of BioC althoug I'm not sure if this is the one you need. You will find it in the Bioc-development site. http://www.bioconductor.org/packages/2.5/data/annotation/html/org.At.t air.db.html Once you have installed it the use with goProfiles should be straightforward. For instance library(goProfiles) library(org.At.tair.db) tab1 <- toTable(org.At.tairENTREZID) genes <- sample(tab1[,1],100) ara.MF.Profiles <-basicProfile(genelist=genes, onto='MF', level=2, orgPackage="org.At.tair.db", empty.cats=FALSE) print(ara.MF.Profiles) will yield a basic profile at level 2 of the MF ontology for a random sample of 100 Arabidopsis genes I have tried it with R-2.10.0 but it should work with R-2.9.z Alex ---------------------------------------------------------------------- ------------------------------- Dr. Alex S?nchez. Associate professor. Statistics Department. University of Barcelona. Facultat de Biologia UB. Avda Diagonal 645. 08028 Barcelona. Spain asanchez_at_ub.edu Statistics and Bioinformatics Unit Institut de Recerca. Hospital Universitari Vall 'Hebron Passeig Vall d'Hebron 112-119. 08034 Barcelona asanchez_at_ir.vhebron.net ---------------------------------------------------------------------- ------------------------------ ----- Original Message ----- From: "Johannes Hanson" <s.j.hanson@uu.nl> To: "Alex Sanchez" <asanchez at="" ub.edu=""> Sent: Wednesday, August 19, 2009 11:25 AM Subject: Re: [BioC] How to compare GO results Thanks, THis software is answering most of the questions I have - excellent. However I am working with Arabidopsis data (Affymetrix ATH1 slides). As far as I understand the annotations files are not availible for arabidopsis. Is that something that can be solved no or in the future releases. Johannes On 15 Sep 2009, at 10:03, Alex Sanchez wrote: > Hi Johannes > > There are several packages in BioC that you can use to do GO analysis or > to mine GO analysis results. > You can find most, if not all, of them in the Bioconductor GO task view. > (Curiously topGO is not there). > > http://www.bioconductor.org/packages/release/GO.html > > I am the maintainer of the 'goProfiles' package which is, in principle, > developed to answer the type of question you pose. > If you have any questions about its use I'll be pleased to try to help > you at answering them. > > best > > Alex S?nchez > > .................................................................... ........................................... > Dr. Alex S?nchez. > Associate Professor. Statistics Department. University of Barcelona. > Facultat de Biologia UB. Avda Diagonal 645. 08028 Barcelona. Spain > asanchez_at_ub.edu > Statistics and Bioinformatics Unit > Institut de Recerca. Hospital Universitari Vall 'Hebron > Passeig Vall d'Hebron 112-119. 08034 Barcelona > asanchez_at_ir.vhebron.net > .................................................................... ........................................... > > > > > > >> Message: 17 >> Date: Wed, 12 Aug 2009 09:10:56 +0200 >> From: Johannes Hanson <s.j.hanson at="" uu.nl=""> >> Subject: [BioC] How to compare GO results >> To: bioconductor at stat.math.ethz.ch >> Message-ID: <02C00589-9261-449C-B8A4-D7C8C4986E72 at uu.nl> >> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes >> >> Dear all, >> >> This might not be a specific BioConductor question. However, >> BioConductor is my major tool and therefore I ask the question here. >> If not appropriate please ignore. >> >> I am analyzing array data from a set of mutants all showing the same >> phenotype. The idea was of course to find genes similarly expressed in >> all mutants to get closer to the regulatory mechanism of the phenotype. >> Surprisingly there was nearly no overlap among the differentially >> expressed genes. The suggestion was raised that the genes might encode >> similar functions. I therefore conducted GO analysis (using topGO and >> other packages). Still very little obvious overlap but how to quantify >> it? How to determine the degree of similarity between different GO >> results? >> >> All suggestions are welcome. >> >> Johannes Hanson >> Molecular Plant Physiology, Utrecht University >> >> >