Using Pairwise Ordering Preferences to Estimate Causal Effects in Gene Expression from a Mixture of Observational and Intervention Experiments

摘要

Animportantstepinsystemsbiologyistoimproveourknowledgeofhowgenescausallyinteractwithoneanother.Afewapproacheshavebeenproposedfortheestimationofcausaleffectsamonggenes,eitherbasedonobservationaldataaloneorrequiringaverypreciseinterventiondesignwithoneknock-outexperimentforeachgene.Werecentlysuggestedamoreflexiblealgorithm,usingaMarkovchainMonteCarloalgorithmandtheMallowsrankingmodel,thatcananalyzeanyinterventiondesign,includingpartialormultipleknock-outs,usingtheframeworkofGaussianBayesiannetworks.Wepreviouslydemonstratedthesuperiorperformanceofthisalgorithmincomparisontoalternativemethods,althoughitcanbecomputationallyexpensivetoimplement.TheaimofthispaperistoproposeanalternativeapproachtakingadvantageofnodepairorderingpreferencestosampletheposteriordistributionaccordingtotheBabington-Smithrankingdistribution.Thisnovelalgorithmproved,bothinasimulationstudyandontheDREAM4challengedata,toattainestimationofthecausaleffectsasaccurateastheMCMC-Mallowsapproachwithahighlyimprovedcomputationalefficiency,beingatleast100timesfaster.WealsotestedouralgorithmontheRosettaCompendiumdatasetwithmorecontrastedresults.Weneverthelessanticipatethatournewapproachmightbeveryusefulforpracticalbiologicalapplications.