A-New-Dynamic-Bayesian-Network-(DBN)-Approach-for-Identifying-一种新的动态贝叶斯网络识别方法课件.ppt

1、A New Dynamic Bayesian Network(DBN)Approach for Identifying Gene Regulatory Networks from Time Course Microarray DataJim VallandinghamBy Min Zou and Suzanne ConzenDynamic Bayesian Networks(DBN)For modeling time-series data Such as microarray data capture the fact that time flows forward Interested i

2、n how genes regulate each other over timeDynamic Bayesian Networks(DBN)Dynamic Bayesian Networks(DBN)Time Problems with DBNs1.Lack a way to determine biologically relevant transcriptional time lagCurrent methods assume same time lag for all potential regulator-target pairsResults in low accuracy of

3、predicting gene relationships2.Excessive computational costPrevents use of DBNs with large scale datasetsNew DBN Method Improvements1.Determine biologically relevant transcriptional time lag Look at initial regulation of regulator and potential target to determine time lag Analyzed for each relation

4、ship Will improve relation predictions 2.Reduce computational cost Only consider genes(up/down)regulated before or at the same time as potential target Reduces search space Reduces costGeneral Outline of MethodHypothetical Example Used to illustrate novel DBN approach 4 hypothetical genes:A D 6 time

5、 points T1 T6 Evenly spaced:indicative of actual data sets.Process broken into 3 major stepsHypothetical Example Step 1:Selection of potential regulators for each gene First,determine time points of changes in expression for each gene Used Thresholds to determine when regulation has occurred 1.2 fol

6、d up-regulation 0.7 fold down-regulation Find Potential RegulatorsPick only genes with earlier or simultaneous changes as regulator candidatesUsed to reduce number of nodes considered Hypothetical ExampleInitial up-regulation at T4Up-regulation thresholdDown-regulation thresholdDynamic Expression pr

7、ofile for Gene DHypothetical ExampleDynamic Expression profiles for Genes A&DPossible regulator of DHypothetical ExampleHypothetical Example Step 2:Estimation of biologically relevant transcriptional time lag Time between expression changes of potential regulator and target genes represents a biolog

8、ically relevant time period Can vary from 0(simultaneous)to many steps Using this time period should result in an increase of correct relationships Hypothetical Example Step 2,cont:Looking at D as target gene and A-C as potential regulators A:2 time units B:2 time units C:1 time unit Group potential

9、 regulators based on time lagA&BCHypothetical Examplet=two time units Hypothetical ExampleHypothetical Example Step 3:Gene regulatory network modeling Use DBN to predict gene regulatory network For DBN variables:Use 2 if expression level is equal to or higher than average expression level over all t

10、ime points Use 1 if expression level is lower than average level Focus of DBN is to predict correlation,not expression value for any given point Hypothetical Example Step 3,cont:Generate subgroups of groups of potential regulators based on user defined minimum and maximum regulators For Hypothetical

11、 Example:Subsets for group A&B A,B,A,B Assuming maximum 2,minimum 1 Subset for group C CHypothetical ExampleStep 3,cont:For each subset,using transcriptional time lag from step 2 to organize expression data into NxM matrixN:number of potential regulators+targetT:number of time points from original s

12、amplingt:estimated transcriptional time lag From step 2M:number of time points in the data matrix=T t Hypothetical ExampleStep 3,cont:Expression value of potential regulators at time Tn are aligned with expression value of the target gene at time Tn+t t may vary for the different expression data mat

13、ricesHypothetical Example1=below average2=at or above averageHypothetical ExampleStep 3,cont:Matrix used to find conditional probabilities of the expression the of target gene in relation to its potential regulator gene(s)Pick subset of potential regulator(s)with highest log marginal likelihood scor

14、e as the estimated regulator(s)Hypothetical ExampleConditional probabilitiesTarget DRegulator AGeneT1T2T3T4A1222D1221Discrete Expression ValuesBiological Experimentation Comparison between their method and standard DBN from Murphys BNT Toolkit Used Chous yeast cell cycle data as input Large number o

15、f time steps(16)and small time intervals(10 min)Previously established relationships used to confirm accuracy.Biological ExperimentationPartial view of previously known regulator-target pairsSimon,I.,Barnett,J.,et al(2019)Serial regulation of transcriptional regulators in the yeast cell cycle.Cell,1

16、06,697708.Biological ExperimentationTwo experiments performed:1.Prior knowledge usedUsed 9 transcription factors to be considered regulators of the 116 genes totalLooked for the targets of these TFs2.No prior knowledge usedAllowed any TF to target pairingBiological Experimentation Metrics Correctly

17、Identified Relationships Association matches previously found pathways Misdirected Association in reverse order of known relationship Specificity Percentage of correctly predicted known gene relationships/total number of predicted gene relationships Computational time Time required to generate DBNEx

18、periment 1 ResultsNetworkCorrectly Identified RelationshipsMisdirected RelationshipsSpecificity(%)Computational Time(s)Novel Method4614010Standard 18411609 TFs considered as regulatorsExperiment 1 Results All misdirected relationships of BNT corrected in new approach.1 Misdirection in new approach d

19、ue to earlier initial up regulation of target rather than regulator For 70%of known yeast interactions the regulator expression change is earlier or simultaneous with targetExperiment 1 Results Using estimated transcriptional time lags gave all relationships found only in new method stronger statist

20、ical correlations Indicates that a biologically relevant and variable time lag is better suited to finding associations than static time step determined by sampling rate 8 uniquely identified by BNT 3 had better correlation at one time step(10 min)rather than the estimated(0 min)5 had earlier expres

21、sion changes of target than regulatorsExperiment 2 ResultsNetworkCorrectly Identified RelationshipsMisdirected RelationshipsSpecificity(%)Computational TimeNovel Method1731015 minStandard 8738 hoursNo previous knowledge consideredExperiment 2 Results 12 correct relations found in new approach and no

22、t BNT 10 of these found because of the use of biological relevant transcriptional time lagSummary New method does increase relationship prediction over standard method in BNT Also significantly reduces computational time Requires min&max number of regulators Requires thresholds for up and down regulation Pre/simultaneous regulator assumption only holds true for 70%of the relationships Missing relevant relationshipsRelationships Found Exp 1