Physiological ProfilingMathematical Integration of Genotypes & PhenotypesPeter J. Tonellato, Ph.D.Director, Bioinformatics Research CenterMedical College of WisconsinPhysiological ProfilingA methodology to discover (data mining) and apply(stratification) empirical or mechanistic based relationships within or between phenotypes and genetic data. Can be equally applied to model organism and human clinical data.PhenotypesMultivariateAnalysisPost-ProcessAcquiredDataMathModelsGenotypesDerivedPhenotypeAnalysisPhysiologicalInformationPhenotypesLinked toGenomePhysiology Linked to GenomeGenomePhysiological ProfilingFour Steps:1.) Collect and maintain functional measures especially for well described genomic backgrounds. (PhysGen)2.) Analyze physiological relationships (both empirical and model based approaches).3.) Analyze phenotype – genetic relationships (s.a. linkage analysis or other predictive modeling).4.) Apply stratification (cluster identification – cluster grouping) to detect primary and secondary relationships between measures.Physiological ProfilingFour Steps:1.) Collect and maintain functional measures especially for well described genomic backgrounds. (PhysGen)2.) Analyze physiological relationships (both empirical and model based approaches).3.) Analyze phenotype – genetic relationships (s.a. linkage analysis or other predictive modeling).4.) Apply stratification (cluster identification – cluster grouping) to detect primary and secondary relationships between measures.Aquired and Derived PhenotypesAcquired DataBP Recording (4 hr.), ?σµ,PhenotypePUT A RAT PICTURE HEREKidney Weight Kidney WeightMABP (2 min)ABP (10-15 min.)RBF (10-15 min.)MRBF (2 min)RBFMABPRVR =2.) Analyze physiological relationships (both empirical and model based approaches).• Capture first order physiological phenomenae.g. correlation of paired measures.• Model development can provide framework for hypothesis development and testinge.g. first order biochemical network (MM)• Create mechanistic(as opposed to empirical) relationships between phenotypese.g. Model of BP variation.2.) Analyze physiological relationships (both empirical and model based approaches).• Capture first order physiological phenomenae.g. correlation of paired measures.GET ONE OF THE SPECIFIC DATA SETS (TWO PHENOTYPES) Between BN and BN.SS.18 and discuss relationship2.) Analyze physiological relationships (both empirical and model based approaches).• Capture first order physiological phenomenae.g. correlation of paired measures.• Model development can provide framework for hypothesis development and testinge.g. first order biochemical network (MM)• Create mechanistic(as opposed to empirical) relationships between phenotypese.g. Model of BP variation.2.) Analyze physiological relationships (both empirical and model based approaches).• Capture first order physiological phenomenae.g. correlation of paired measures.• Model development can provide framework for hypothesis development and testinge.g. first order biochemical network (MM)• Create mechanistic(as opposed to empirical) relationships between phenotypese.g. Model of BP variation.MAP from Hypertensive, Normal, and SubtypesecondsMAP (mmHg)Physiological ProfilingFour Steps:1.) Collect and maintain functional measures especially for well described genomic backgrounds. (PhysGen)2.) Analyze physiological relationships (both empirical and model based approaches).3.) Analyze phenotype – genetic relationships (s.a. linkage analysis or other predictive modeling).4.) Apply stratification (cluster identification – cluster grouping) to detect primary and secondary relationships between measures.100 150 200 25002468100 150 200 2500510152025systolic blood pressure(mmHg)freq.freq.BN SSF2High Salt Blood PressuresecondssecondssecondssecondsmmHgmmHgmmHgmmHgRat to Human: Model DiscoveryRH G750 272.0741 251.0248 123.9360 152.2159 87.7155 81.675.4 48.8cR cMHumanGLUE02040608010014012012/116UniGenesRH619 440189 312 168 130 13 RATQTLsGcMcRHuman to Rat: Model Creation HumanRATUniGenesQTLsRH G750 272.0741 251.0248 123.9360 152.2159 87.7155 81.675.4 48.8020406080100140120RH619 440189 312 168 130 13 GGLUEcM cR cR cM12/116Physiological ProfilingFour Steps:1.) Collect and maintain functional measures especially for well described genomic backgrounds. (PhysGen)2.) Analyze physiological relationships (both empirical and model based approaches).3.) Analyze phenotype – genetic relationships (s.a. linkage analysis or other predictive modeling).4.) Apply stratification (cluster identification – cluster grouping) to detect primary and secondary relationships between measures.Brown NorwayColorized correlation matrix of 86 phenotypes.Linearly ordered and grouped by prescribed physiological relationships.F2 vs Segregated F2(Low responders to salt loading)+1 0 -1F2_lt+1 0 -1F2 vs Segregated F2(Allele for Salt Sensitive trait)Data Mining:Interrogate the population to identify subpopulations of interests.+1 0 -1F2 vs Segregated F2(heterozygotes)Colorized correlation matrix of 86 phenotypes.Linearly ordered and grouped by prescribed physiological relationships.Data Mining:Interrogate the population to identify subpopulations of interests.AB C D F GEClusters of BN correlationF2F2SSBNPhenotypes grouped by BN clustersConclusions• Functional relationships should be fully explored: Modeling, Empirical Testing and Mapping• Combination of analysis to develop integrated perspective: Function, Structure, and interaction• Physiological profiling is a discovery process that should lead to hypothesis concerning physiological function and the corresponding genetic basisAcknowledgementsBioinformatics Research CenterSimon Twigger Zhitao Wang Jed MathisMary Shimoyama Dean Pasko Jessica GinsterDan Chen Peter Long Mike ThomasHaiping Xia Wiehong Jin Hai YuSam Cheng Nan JiangRoumyana Kirova Nancy SobczekDept. of PhysiologyAllen W. CowleyAndrew S. Greeneand A cast of dozensHuman and Molecular GeneticsHoward Jacob Rebecca MajewskiAnne Kwitek-Black Li GuoYashiro Kita Marcelo NobregaMasahide Shiozawa Pierre DumasUlrich Broeckel Sheri JeneChristian-Albrechts-UniversitätKiel Institut für PharmakologieMonika