microarray-ontol-digest Thursday, May 30 2002 Volume 01 : Number 025 ---------------------------------------------------------------------- Date: Sun, 19 May 2002 17:39:07 +0200 (MET DST) From: Kurt Fellenberg Subject: Re: [microarray-ontol] proposed ontology classes Dear all, > BiologicalFactor > Definition: Factors that are varied that are not methodological, so > things that relate to the biomaterial and its treatment and growth. > Subclasses: GeneticVariationFactor, StressFactor, TemperatureFactor, > TimeCourseFactor would like to propose (an) additional subclass(es) for factors not fitting to the ones listed. An example: A transgene was transfected into WT yeast cells under the control of a galactose inducible promoter. Because galactose affects the transcription of some genes, one has to subtract the effects of galactose from those of the transgene. Experimental conditions: 1. WT without galactose (replaced by glucose as a carbon source) 2. WT with galactose 3. transgenic strain without 4. transgenic strain with galactose (complete description at http://www.dkfz-heidelberg.de/tbi/services/mchips/cdc14.html) Galactose is replaced by another carbon source, thus the cells are not starved or stressed, so StressFactor would not apply. It may be annotated e.g. as "NutrientFactor" or simply as "OtherEnvironmentalFactor". Might be more realistic to use the latter term or the list (of subclasses) is likely to become a long one: * other (temporary) additives to the growth medium not stressing the cells, such as hormones, calcium, ... * liquid or solid culture * growth phase (for yeast, Neurospora crassa, ...) * aerob / anaerob growth * light source, period, intensity for plants * tumor stage / tumor grade (or would it be a "GeneticVariation"?) * the host for e.g. Trypanosoma brucei (in mouse, fly or grown in vitro) * all kinds of phenotypes observed e.g. for mice (incl. morphological abnormality, spatial learning, ...) that may correlate with transcription patterns ... (more parameters at http://www.dkfz-heidelberg.de/tbi/services/mchips) On the other hand, it might be boring to be too realistic :-). Depends on the level of detail desirable for modeling complex biological samples. What do you think? All the best, Kurt - ------------------------------------------------------------------------------- Kurt Fellenberg German Cancer Research Center Tel: + 49 + 6221 / 42-2718 Theoretical Bioinformatics (H0300) Fax: + 49 + 6221 / 42-2849 Im Neuenheimer Feld 280 email: K.Fellenberg@DKFZ-Heidelberg.de D-69120 Heidelberg, Germany URL: http://www.DKFZ-Heidelberg.de/tbi/people/fellenberg http://www.dkfz-heidelberg.de/tbi/services/mchips - ------------------------------------------------------------------------------- > > Dear Group, > Below is a proposal for classes to add to the ontology addressing > experimental design. These represent an important type of annotation > typically entered through forms as with biomaterial descriptions. These > classes are based on: > 1. MIAME 1.1 and MAGE Jan 2002 Gene Expression RFP document by Michael > Miller (see pages 72-75). > 2. Helen Parkinson's initial expfactor ontology > 3. comments from John Potter (Fred Hutchinson Cancer Research Center) at > Oncogenomics 2002 regarding the distinction between factors that are > controlled during an experiment and factors which are observed or > measured during a survey. > > Please note that this is just a start. I will add these to the ontology > next weekend following any discussion. > > Cheers, > Chris > > New proposed classes for describing the experiment/study (subclasses are > indented): > Study > StudyDesign > ReplicateDesign > NormalizationDesign > QualityControlDesign > StudyFactor > HistoryFactor > BiologicalFactor > GeneticVariationFactor > StressFactor > TemperatureFactor > TimeCourseFactor > MethodologicalFactor > HardwareVariation > SoftwareVariation > ProtocolVariation > ### > Study > Definition: The complete set of assays and their descriptions performed > as a study for a common purpose. In general, a study will be equivalent > to a publication and have a single submitter. (associations to analysis > results, assays, and data will be added later). > Properties: > has_submitter (Contact) > has_publication (BibliographicReference) > has_URI (URI) > has_study_design(StudyDesign) > number_of_assays > ### > StudyDesign > Definition: The organizing principles of the study including the > relationships between assays and steps taken to interpret the data. Used > to organize the assays into groups. > Subclasses: ReplicateDesign, NormalizationDesign, QualityControlDesign > > Properties: > study_group_id > has_study_factor(StudyFactor) > has_description > has_study_type(one of: normal vs diseased, treated vs unstreated, > time course, tiling) > ### > ReplicateDesign > Definition: Groups of assays that represent replicates > Properties: > replicate_type (one_of: technical, biological) > ### > NormalizationDesign > Definition: Groups of assays whose data have been processed to be > directly comparable > Properties: > normalization_type (one-of: global lowess, print-tip global lowess, > linrear regression) > ### > QualityControlDesign > Definition: Groups of assays that are related as part of quality control. > Properties: > control_type (one-of: dye-swap, common reference) > ### > StudyFactor > Definition: The factors in the study that are experimental parameters > or regarded as influencing the study results. Factor group ID will be > used to link assays to the factors associated with them. > Sublcasses: BiologicalFactor, MethodologicalFactor > Properties: > has_factor_value (Measurement) > factor_group_id > ### > HistoryFactor > Definition: Factors that are measured or observed parts of the study but > not induced or under the control of the experiment. > Properties: > has_environmental_history(EnvironmentalHistory) > ### > BiologicalFactor > Definition: Factors that are varied that are not methodological, so > things that relate to the biomaterial and its treatment and growth. > Subclasses: GeneticVariationFactor, StressFactor, TemperatureFactor, > TimeCourseFactor > Properties: > has_treatment(Treatment) > ### > MethodologicalFactor > Definition: Methodological variations. Not things that relate to the > biomaterial but things that are expressed as protocols for example. Only > needed if relevant for experimental design. So if the point of the > experiment was to compare data across two different scanners, array > designs, or between teo (or more) different hybridization chambers then > these should be expressed as experimental factors (MAGE terms) and can > be given factor values. These should be expressed as instances in this > ontology so the name of a hybridization chamber could be an instance > here. > Subclasses: HardwareVariation, SoftwareVariation, ProtocolVariation > ### > GeneticVariationFactor > Definition: Describes genetically modified organisms. Does not refer to > natural genetic variation > ### > StressFactor > Definition: Degree of stress induced. > ### > TemperatureFactor > Definition: Temperature is varied. > ### > TimeCourseFactor > Definition: Time course refers to an experiment which has multiple time > points. Cell cycle is an instance of time course. > ### > HardwareVariation > Definition: Hardware variations that are part of experiment design. > Properties: > has_hardware(Hardware) > ### > SoftwareVariation > Definition: Software variations that are part of experiment design. > Properties: > has_software(Software) > ### > ProtocolVariation > Definition: Protocol variations that are part of experiment design. > Properties: > has_protocol(Protocol) > > ------------------------------ Date: Tue, 21 May 2002 10:08:12 -0400 From: Chris Stoeckert Subject: Re: [microarray-ontol] proposed ontology classes Hi Kurt, The subclasses for BiologicalFactor are meant to capture a common type of treatment (note that they inherit a link to Treatment from BiologicalFactor). Current subclasses of Treatment are BehavioralStimulus, CompoundBasedTreatment, ContaminantOrganism, GrowthCondition, Infection, Modification, Preservation, and Starvation. GrowthCondition and Modification have additional subclasses including Nutrients for GrowthCondition. I'm thinking we should try to use these Treatments to guide structuring of BiologicalFactors. For your example we could create a BiologicalFactor subclass GrowthConditionFactor which has a subclass NutrientFactor. We might even constrain NutrientFactor to only associate with Nutrients. I'm not sure that this is the best way to do this so I am open to other suggestions. GIven that I have basically defined BiologicalFactor as a Treatment we may need to rethink TimeCourseFactor (how do you treat with time?). Cheers, Chris > would like to propose (an) additional subclass(es) for factors not > fitting > to the ones listed. An example: > > A transgene was transfected into WT yeast cells under the control of a > galactose inducible promoter. Because galactose affects the > transcription of > some genes, one has to subtract the effects of galactose from those of > the > transgene. Experimental conditions: > > 1. WT without galactose (replaced by glucose as a carbon source) > 2. WT with galactose > 3. transgenic strain without > 4. transgenic strain with galactose > > (complete description at http://www.dkfz- > heidelberg.de/tbi/services/mchips/cdc14.html) > > Galactose is replaced by another carbon source, thus the cells are not > starved or stressed, so StressFactor would not apply. It may be > annotated > e.g. as "NutrientFactor" or simply as "OtherEnvironmentalFactor". > > Might be more realistic to use the latter term or the list (of > subclasses) > is likely to become a long one: > > * other (temporary) additives to the growth medium not stressing the > cells, > such as hormones, calcium, ... > * liquid or solid culture > * growth phase (for yeast, Neurospora crassa, ...) > * aerob / anaerob growth > * light source, period, intensity for plants > * tumor stage / tumor grade (or would it be a "GeneticVariation"?) > * the host for e.g. Trypanosoma brucei (in mouse, fly or grown in vitro) > * all kinds of phenotypes observed e.g. for mice > (incl. morphological abnormality, spatial learning, ...) that may > correlate > with transcription patterns > ... (more parameters at http://www.dkfz- > heidelberg.de/tbi/services/mchips) > > On the other hand, it might be boring to be too realistic :-). Depends > on > the level of detail desirable for modeling complex biological samples. > What > do you think? > > All the best, > Kurt > > ------------------------------------------------------------------------------- > Kurt Fellenberg > German Cancer Research Center Tel: + 49 + 6221 / 42-2718 > Theoretical Bioinformatics (H0300) Fax: + 49 + 6221 / 42-2849 > Im Neuenheimer Feld 280 email: > K.Fellenberg@DKFZ-Heidelberg.de > D-69120 Heidelberg, Germany > > URL: http://www.DKFZ-Heidelberg.de/tbi/people/fellenberg > http://www.dkfz-heidelberg.de/tbi/services/mchips > ------------------------------------------------------------------------------- > >> >> Dear Group, >> Below is a proposal for classes to add to the ontology addressing >> experimental design. These represent an important type of annotation >> typically entered through forms as with biomaterial descriptions. These >> classes are based on: >> 1. MIAME 1.1 and MAGE Jan 2002 Gene Expression RFP document by Michael >> Miller (see pages 72-75). >> 2. Helen Parkinson's initial expfactor ontology >> 3. comments from John Potter (Fred Hutchinson Cancer Research Center) >> at >> Oncogenomics 2002 regarding the distinction between factors that are >> controlled during an experiment and factors which are observed or >> measured during a survey. >> >> Please note that this is just a start. I will add these to the ontology >> next weekend following any discussion. >> >> Cheers, >> Chris >> >> New proposed classes for describing the experiment/study (subclasses >> are >> indented): >> Study >> StudyDesign >> ReplicateDesign >> NormalizationDesign >> QualityControlDesign >> StudyFactor >> HistoryFactor >> BiologicalFactor >> GeneticVariationFactor >> StressFactor >> TemperatureFactor >> TimeCourseFactor >> MethodologicalFactor >> HardwareVariation >> SoftwareVariation >> ProtocolVariation >> ### >> Study >> Definition: The complete set of assays and their descriptions performed >> as a study for a common purpose. In general, a study will be equivalent >> to a publication and have a single submitter. (associations to analysis >> results, assays, and data will be added later). >> Properties: >> has_submitter (Contact) >> has_publication (BibliographicReference) >> has_URI (URI) >> has_study_design(StudyDesign) >> number_of_assays >> ### >> StudyDesign >> Definition: The organizing principles of the study including the >> relationships between assays and steps taken to interpret the data. >> Used >> to organize the assays into groups. >> Subclasses: ReplicateDesign, NormalizationDesign, QualityControlDesign >> >> Properties: >> study_group_id >> has_study_factor(StudyFactor) >> has_description >> has_study_type(one of: normal vs diseased, treated vs unstreated, >> time course, tiling) >> ### >> ReplicateDesign >> Definition: Groups of assays that represent replicates >> Properties: >> replicate_type (one_of: technical, biological) >> ### >> NormalizationDesign >> Definition: Groups of assays whose data have been processed to be >> directly comparable >> Properties: >> normalization_type (one-of: global lowess, print-tip global lowess, >> linrear regression) >> ### >> QualityControlDesign >> Definition: Groups of assays that are related as part of quality >> control. >> Properties: >> control_type (one-of: dye-swap, common reference) >> ### >> StudyFactor >> Definition: The factors in the study that are experimental parameters >> or regarded as influencing the study results. Factor group ID will be >> used to link assays to the factors associated with them. >> Sublcasses: BiologicalFactor, MethodologicalFactor >> Properties: >> has_factor_value (Measurement) >> factor_group_id >> ### >> HistoryFactor >> Definition: Factors that are measured or observed parts of the study >> but >> not induced or under the control of the experiment. >> Properties: >> has_environmental_history(EnvironmentalHistory) >> ### >> BiologicalFactor >> Definition: Factors that are varied that are not methodological, so >> things that relate to the biomaterial and its treatment and growth. >> Subclasses: GeneticVariationFactor, StressFactor, TemperatureFactor, >> TimeCourseFactor >> Properties: >> has_treatment(Treatment) >> ### >> MethodologicalFactor >> Definition: Methodological variations. Not things that relate to the >> biomaterial but things that are expressed as protocols for example. >> Only >> needed if relevant for experimental design. So if the point of the >> experiment was to compare data across two different scanners, array >> designs, or between teo (or more) different hybridization chambers then >> these should be expressed as experimental factors (MAGE terms) and can >> be given factor values. These should be expressed as instances in this >> ontology so the name of a hybridization chamber could be an instance >> here. >> Subclasses: HardwareVariation, SoftwareVariation, ProtocolVariation >> ### >> GeneticVariationFactor >> Definition: Describes genetically modified organisms. Does not refer to >> natural genetic variation >> ### >> StressFactor >> Definition: Degree of stress induced. >> ### >> TemperatureFactor >> Definition: Temperature is varied. >> ### >> TimeCourseFactor >> Definition: Time course refers to an experiment which has multiple time >> points. Cell cycle is an instance of time course. >> ### >> HardwareVariation >> Definition: Hardware variations that are part of experiment design. >> Properties: >> has_hardware(Hardware) >> ### >> SoftwareVariation >> Definition: Software variations that are part of experiment design. >> Properties: >> has_software(Software) >> ### >> ProtocolVariation >> Definition: Protocol variations that are part of experiment design. >> Properties: >> has_protocol(Protocol) >> >> > ------------------------------ Date: Wed, 22 May 2002 16:13:12 +0200 (MET DST) From: Kurt Fellenberg Subject: Re: [microarray-ontol] proposed ontology classes Hi Chris, > The subclasses for BiologicalFactor are meant to capture a common type > of treatment (note that they inherit a link to Treatment from > BiologicalFactor). Current subclasses of Treatment are > BehavioralStimulus, CompoundBasedTreatment, ContaminantOrganism, > GrowthCondition, Infection, Modification, Preservation, and Starvation. > GrowthCondition and Modification have additional subclasses including > Nutrients for GrowthCondition. I'm thinking we should try to use these > Treatments to guide structuring of BiologicalFactors. For your example > we could create a BiologicalFactor subclass GrowthConditionFactor which > has a subclass NutrientFactor. I see. > We might even constrain NutrientFactor to > only associate with Nutrients. I'm not sure that this is the best way to > do this so I am open to other suggestions. GIven that I have basically > defined BiologicalFactor as a Treatment we may need to rethink > TimeCourseFactor (how do you treat with time?). A) In principle, one could study adaption of cells to any treatment over time. B) Another type of experiment might be a treatment that is varied over time. Alteration of too many parameters in an experiment complicates interpretation of the outcome. B) may be problematic in that sense. I don't know if it is needed at the moment. But one might want to prepare for B) as well. Cheers, Kurt ------------------------------ Date: Mon, 27 May 2002 13:37:26 +0100 From: Helen Parkinson Subject: [microarray-ontol] Tissue Type/Organ vs Organism Part Hi, a quick look around a few websites indicates that many people are using the concept tissue type to describe expt, eg EpoDB and maxDB, (thanks to Dave Hancock for pointing flagging this). There are also ontologies that are in development for tissue type. The MGED ontology has OrganismPart, I think this was chosen so that we didn't have any species dependent distinction, but I can think of cases where we might want to specify both organ and tissue type. In any case OrganismPart is currently undefined. eg cardiac muscle Organ: heart Tissue type: cardiac muscle Could we refine OrganismPart so that it gets subclasses with the definitions (from Ox.Dict.Mol.Biol.) as follows: %BiosourceOntologyEntry %OrganismPart %Organ - any part of the body of a multicellular organism that is adapted and/or specialised for the performance of one or more vital functions. %Tissuetype - any collection of cells that is organised to perform one or more specific functions this splitting could allow the description of more complex terms in a cleaner way? We could either just lose the OrganismPart and create these two new classes or, just make them subclasses, and to go one step further, if we keep OrganismPart, then CellType and TargetCellType are also organism parts so should these also become subclasses? comments anyone? thanks Helen ------------------------------ Date: Mon, 27 May 2002 13:44:35 +0100 From: Helen Parkinson Subject: [microarray-ontol] MGED definitions Hi all, apologies a correction, I have just loaded biomaterial14.rdfs into Oiled and somehow it's lost all the definitions, so in fact OrganismPart is defined as: "The part of the organism's anatomy from which the biomaterial was derived" Does anyone else have the same problem with biomaterial14.rdfs? regards Helen ------------------------------ Date: Mon, 27 May 2002 10:19:22 -0400 From: Chris Stoeckert Subject: [microarray-ontol] Re: MGED definitions Hi Helen, I ran into problems going between OILed 2.2 and 3.4. Which version are you using? (biomaterial14.rdfs was done in 3.4). Have you tried the DAML file? Cheers, Chris On Monday, May 27, 2002, at 08:44 AM, Helen Parkinson wrote: > Hi all, apologies > > a correction, > > I have just loaded biomaterial14.rdfs into Oiled and > somehow it's lost all the definitions, so in fact > OrganismPart is defined as: > > "The part of the organism's anatomy from which the > biomaterial was derived" > > Does anyone else have the same problem with > biomaterial14.rdfs? > > regards > > Helen > ------------------------------ Date: Mon, 27 May 2002 21:48:03 -0400 From: Chris Stoeckert Subject: Re: [microarray-ontol] Tissue Type/Organ vs Organism Part > Date: Mon, 27 May 2002 20:08:58 -0400 > From: Bruce Aronow > X-Mailer: Mozilla 4.77C-CCK-MCD {C-UDP; EBM-APPLE} (Macintosh; U; PPC) > X-Accept-Language: en > MIME-Version: 1.0 > To: Helen Parkinson > CC: ontol > Subject: Re: [microarray-ontol] Tissue Type/Organ vs Organism Part > References: <3CF22886.7B306096@ebi.ac.uk> > Content-Type: text/plain; charset=us-ascii; x-mac-type="54455854"; > x-mac-creator="4D4F5353" > Content-Transfer-Encoding: 7bit > > Hi Helen, > > Between Organism and Organ I would want to have info on :: > > AnatomicRegion.. probably same as OrganismPart (eg.. drosophila head, > antero-ventral embryo quadrant, neck, .. and stuff like that..) > > System (physiologic .. eg nervous, gastrointestinal, endocrine, > muscular, skeletal.. etc) > > below Organ I would have CellType (eg... fractionated T-cells, > endothelial cells, etc.) > > best, > Bruce > > > > Helen Parkinson wrote: > >> Hi, >> >> a quick look around a few websites indicates that many >> people are using the concept tissue type to describe expt, >> eg EpoDB and maxDB, (thanks to Dave Hancock for pointing >> flagging this). There are also ontologies that are in >> development for tissue type. >> >> The MGED ontology has OrganismPart, I think this was chosen >> so that we didn't have any species dependent distinction, >> but I can think of cases where we might want to specify both >> organ and tissue type. In any case OrganismPart is currently >> undefined. >> >> eg cardiac muscle >> >> Organ: heart >> Tissue type: cardiac muscle >> >> Could we refine OrganismPart so that it gets subclasses with >> the definitions (from Ox.Dict.Mol.Biol.) as follows: >> >> >> %BiosourceOntologyEntry >> %OrganismPart >> %Organ - any part of the body of a multicellular organism >> that is adapted and/or specialised for the performance of >> one or more vital functions. >> %Tissuetype - any collection of cells that is organised to >> perform one or more specific functions >> >> this splitting could allow the description of more complex >> terms in a cleaner way? We could either just lose the >> OrganismPart and create these two new classes or, just make >> them subclasses, >> >> and to go one step further, if we keep OrganismPart, then >> CellType and TargetCellType are also organism parts so >> should these also become subclasses? >> >> comments anyone? >> >> thanks >> >> Helen > ------------------------------ Date: Tue, 28 May 2002 16:10:59 +0100 From: Helen Parkinson Subject: [microarray-ontol] cancer disease descriptions Hi Chris and John et al., I have just received a MAGE-ML sample submission from John Yost NCI's annotation tool/database, this has the following OntologyEntries: currently we have %BiosourceOntologyEntry %DiseaseState %ClinicalInformation, are these adequate? The tissue type one is already under discussion, I am supposing that if NCI are generating this type of annotation they want to/already use these. Should these be defined and added to the ontology, and would anyone esp. cancer biologists like to comment? Helen ------------------------------ Date: Wed, 29 May 2002 14:13:45 +0100 From: Helen Parkinson Subject: [microarray-ontol] SOFG conference, 17-20 Nov 2002 The organisers are pleased to announce the first Standards and Ontologies for Functional Genomics (SOFG)*+ conference to be held at Hinxton, 17-20 November 2002. *Wellcome Trust Funding applied for. +An MGED supported conference. Further details: http://www.ebi.ac.uk/microarray/General/Events/SOFG/SOFG.html Confirmed Keynote Speakers: --------------------------- Ken Buetow, NCI, USA Win Hide, SANBI, South Africa Peter Karp, SRI International, USA The goal of this conference is to bring together scientists who are developing standards and ontologies for describing microarray and other high throughput functional genomics experiments. The topics will include GO, anatomy and phenotype ontologies, disease ontologies, chemical compounds and treatments, inter-ontology mapping and tools for ontology development. We look forward to seeing you in November, The Programme Committee: -------------------- Michael Ashburner, University of Cambridge (chair) Cathy Ball, Stanford University, USA. Mike Bittner, NHGRI, USA Alvis Brazma, EMBL-EBI, UK Duncan Davidson, MRC HGU, Edinburgh, UK Liz Ford, EMBL-EBI, UK Midori Harris, EMBL-EBI, UK Victor Markowitz, Gene Logic, USA Martin Ringwald, The Jackson Laboratories, USA Helen Parkinson, EMBL-EBI, UK. John Quackenbush, TIGR, USA Steffen Schulze-Kremer, RZPD, Germany. Paul Spellman, U.C.Berkeley, USA. Robert Stevens, University of Manchester, UK Chris Stoeckert, University of Pennsylvania, USA ------------------------------ Date: Thu, 30 May 09:19:36 2002 -0400 From: "Seetharamulu Peddaiahgari" Subject: [microarray-ontol] remove seetha@origenix.com remove seetha@origenix.com ------------------------------ End of microarray-ontol-digest V1 #25 *************************************