PackedCandidate PF MuonID flags not properly set in PATPackedCandidateProducer? Implications on Type-1 MET re-correction

[nurfikri89]

I have encountered what might be an unintentional wrong set up (?) regarding the MuonID flags of pat::PackedCandidates in MiniAODs. It seems that there are (somewhat rare) instances where pat::PackedCandidate PF muons,
whose isStandAloneMuon() and isGlobalMuon() flags are false but for its original reco::PFCandidate in the parent AOD, the two MuonID flags are true. Technically for the reco::PFCandidate, these flags are retrieved through its muonRef().

This inconsistency can be traced back to how the MuonID flags are set up in PATPackedCandidateProducer:

cmssw/PhysicsTools/PatAlgos/plugins/PATPackedCandidateProducer.cc

Lines 326 to 328 in bbbd522

	if (cand.muonRef().isNonnull()) {
	outPtrP->back().setMuonID(cand.muonRef()->isStandAloneMuon(), cand.muonRef()->isGlobalMuon());
	}

where setMuonID() is called when the reco::PFCandidate has cand.muonRef().isNonnull() and the candidate has a trackRef():

cmssw/PhysicsTools/PatAlgos/plugins/PATPackedCandidateProducer.cc

Line 244 in bbbd522

if (ctrack) {

The instances that I mentioned above are when a reco::PFCandidate muon have a muonRef() but do not have a trackRef() . To fix this inconsistency, we just need to move MuonID flags setting block, to outside of the if (ctrack) block.

One implication for this is the consistency of Type-1 MET corrections for different data-tiers usage (which is the context of my study).

The Type-1 correction procedure requires that PF muon candidates inside a jet are to be removed from the jet raw four-vector if it passes isStandAloneMuon() or isGlobalMuon() flags and this is where the Muon ID flag differences between reco::PFCandidate and pat::PackedCandidate becomes a source of inconsistency between (Type-1 corrected) MET in MiniAOD and the MET re-corrections when they are done during MiniAOD processing.

At the PAT step, the jet cleaning stage (to remove PF muon constituents) loops over reco::PFCandidates while for the NANO step (or any analysis using MiniAODs as input), it loops over pat::PackedCandidates. This leads to situations where the PF muon is removed because it passes either MuonID flags if it is a reco::PFCandidate but it does not get removed when it is a pat::PackedCandidate due to the two ID flags not being set at all in PATPackedCandidateProducer.

[cmsbuild]

cms-bot internal usage

[cmsbuild]

A new Issue was created by @nurfikri89.

@Dr15Jones, @antoniovilela, @makortel, @mandrenguyen, @rappoccio, @sextonkennedy, @smuzaffar can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[makortel]

assign PhysicsTools/PatAlgos

[cmsbuild]

New categories assigned: reconstruction,xpog

@vlimant,@hqucms,@ftorrresd,@jfernan2,@mandrenguyen you have been requested to review this Pull request/Issue and eventually sign? Thanks

[nurfikri89]

Fixed with PR #45818, merged with 14_2_X.

[hqucms]

+1

[jfernan2]

+1

[cmsbuild]

This issue is fully signed and ready to be closed.

[makortel]

@cmsbuild, please close