7.4 Electron Analysis

Complete:
Detailed Review status

Contents

Introduction

The present page deals with the analysis of electron data produced with CMSSW 94X. It shows examples of some simple analysis code reading the standard collection of reconstructed electron objects whose reconstruction is performed by default by CMSSW. For further details about the reconstruction process, have a look at SWGuideElectronRecoSoftware.

Getting Started

Access to computing environment

First make sure that you have access to either cmslpc, or lxplus. You will also need a github account to checkout the code. To check if you’re using bash, tcsh, sh, csh, ksh, or zsh. You can request for a change in shell (cmslpc,lxplus account self-service tool).

You might as well get a proxy now as you will need it later.

voms-proxy-init -voms cms --valid 192:00

If you do not have grid access see LcgAccess

Analysis in CMSSW 94x

Set-up CMSSW Release

Below is a recipe for setting up the analysis for the CMSSW94 release:

cmsrel CMSSW_9_4_4
cd CMSSW_9_4_4/src
cmsenv

Create Ntuples using the ggNtuplizer

Bare ROOT access

At this point you should have set up your computing environment and perhaps have a centrally produced miniAOD ROOT file. the simplest thing to do with an edm ROOT file is to open it with ROOT interpreter (do not care about the numerous warnings about the lacking dictionaries). You can look within the file, for instance look at the tree called "Events". But what we really want to do is to Ntuplize the datasets. An Ntuple is simply a collection of events, where different variables (momentum, Energy, etc..) corresponding to each events are kept. With these Ntuples, one can make plot meaningful distributions and perform an analysis.

To make Ntuples, one can write an EDAnalyzer. However, making a complete one from scratch takes time so there is a standard Ntuplizer for doing photon and electron analysis called the ggNtuplizer.

Create Ntuples using the ggNtuplizer

First we build the ggNtuplizer in the latest CMSSW version it’s available in: https://github.com/cmkuo/ggAnalysis/tree/94X

This step may take some time.

git cms-init 
git cms-merge-topic lsoffi:CMSSW_9_4_0_pre3_TnP 
git cms-merge-topic guitargeek:ElectronID_MVA2017_940pre3 
scram b -j8 
cd $CMSSW_BASE/external/slc6_amd64_gcc630 
git clone https://github.com/lsoffi/RecoEgamma-PhotonIdentification.git data/RecoEgamma/PhotonIdentification/data 
cd data/RecoEgamma/PhotonIdentification/data 
git checkout CMSSW_9_4_0_pre3_TnP 
cd $CMSSW_BASE/external/slc6_amd64_gcc630/ 
git clone https://github.com/lsoffi/RecoEgamma-ElectronIdentification.git data/RecoEgamma/ElectronIdentification/data 
cd data/RecoEgamma/ElectronIdentification/data 
git checkout CMSSW_9_4_0_pre3_TnP 
cd $CMSSW_BASE/src 
git cms-merge-topic cms-egamma:EGM_94X_v1 
cd EgammaAnalysis/ElectronTools/data 
git clone https://github.com/ECALELFS/ScalesSmearings.git 
cd ScalesSmearings/ 
git checkout Run2017_17Nov2017_v1
cd $CMSSW_BASE/src 
git clone https://github.com/cmkuo/HiggsAnalysis.git 
git clone -b 94X https://github.com/cmkuo/ggAnalysis.git 
scram b -j8 

You can now make Ntuples using a configuration file. You can open it with a text editor like vim,

vi ggAnalysis/ggNtuplizer/test/run_data_94X.py #or 
vi ggAnalysis/ggNtuplizer/test/run_mc_94X.py

If you want to learn what the Ntuplize does you can check out,

vi ggAnalysis/ggNtuplizer/plugins/ggNtuplizer_photons.cc.

One dataset we could run on is DYJetsToLL_M-50_TuneCP5_13TeV-amcatnloFXFX-pythia8/MINIAODSIM/94X_mc2017_realistic_v10-v1/00000/005DC030-D3F4-E711-889A-02163E01A62D.root (see CMS Data Aggregation system).

In the run_mc_94X.py configuration file, comment out the original input file and add the file we want to process.

process.source = cms.Source("PoolSource",
                            fileNames = cms.untracked.vstring(
        #'file:/data4/cmkuo/testfiles/DYJetsToLL_M-50_RunIIFall17.root'        
        '/store/mc/RunIIFall17MiniAOD/DYJetsToLL_M-50_TuneCP5_13TeV-amcatnloFXFX-pythia8/MINIAODSIM/94X_mc2017_realistic_v10-v1/00000/005DC030-D3F4-E711-889A-02163E01A62D.root'
        ))

Let's set the process.maxEvents to just run over 10000 events. To execute the process go to the source directory and do cmsRun,

 cd $CMSSW_BASE/src/
cmsRun ggAnalysis/ggNtuplizer/test/run_mc_94X.py

The output should look like this,

	--- egmGsfElectronIDs:cutBasedElectronID-Fall17-94X-V1-loose added to patElectrons
	--- egmGsfElectronIDs:cutBasedElectronID-Fall17-94X-V1-medium added to patElectrons
	--- egmGsfElectronIDs:cutBasedElectronID-Fall17-94X-V1-tight added to patElectrons
	--- egmGsfElectronIDs:cutBasedElectronID-Fall17-94X-V1-veto added to patElectrons
	--- egmGsfElectronIDs:heepElectronID-HEEPV70 added to patElectrons
	--- egmGsfElectronIDs:mvaEleID-Fall17-noIso-V1-wp80 added to patElectrons
	--- egmGsfElectronIDs:mvaEleID-Fall17-noIso-V1-wp90 added to patElectrons
	--- egmGsfElectronIDs:mvaEleID-Fall17-noIso-V1-wpLoose added to patElectrons
	--- egmGsfElectronIDs:mvaEleID-Fall17-iso-V1-wp80 added to patElectrons
	--- egmGsfElectronIDs:mvaEleID-Fall17-iso-V1-wp90 added to patElectrons
	--- egmGsfElectronIDs:mvaEleID-Fall17-iso-V1-wpLoose added to patElectrons
	--- egmPhotonIDs:cutBasedPhotonID-Fall17-94X-V1-loose added to patPhotons
	--- egmPhotonIDs:cutBasedPhotonID-Fall17-94X-V1-medium added to patPhotons
	--- egmPhotonIDs:cutBasedPhotonID-Fall17-94X-V1-tight added to patPhotons
	--- egmPhotonIDs:mvaPhoID-RunIIFall17-v1-wp80 added to patPhotons
	--- egmPhotonIDs:mvaPhoID-RunIIFall17-v1-wp90 added to patPhotons
10-Oct-2018 00:59:58 CDT  Initiating request to open file root://cmsxrootd-site.fnal.gov//store/mc/RunIIFall17MiniAOD/DYJetsToLL_M-50_TuneCP5_13TeV-amcatnloFXFX-pythia8/MINIAODSIM/94X_mc2017_realistic_v10-v1/00000/005DC030-D3F4-E711-889A-02163E01A62D.root
%MSG-w XrdAdaptor:  file_open 10-Oct-2018 00:59:59 CDT pre-events
Data is served from fnal.gov instead of original site T1_US_FNAL
%MSG
10-Oct-2018 01:00:00 CDT  Successfully opened file root://cmsxrootd-site.fnal.gov//store/mc/RunIIFall17MiniAOD/DYJetsToLL_M-50_TuneCP5_13TeV-amcatnloFXFX-pythia8/MINIAODSIM/94X_mc2017_realistic_v10-v1/00000/005DC030-D3F4-E711-889A-02163E01A62D.root
Begin processing the 1st record. Run 1, Event 74714917, LumiSection 42879 at 10-Oct-2018 01:00:23.365 CDT
10-Oct-2018 01:01:20 CDT  Closed file root://cmsxrootd-site.fnal.gov//store/mc/RunIIFall17MiniAOD/DYJetsToLL_M-50_TuneCP5_13TeV-amcatnloFXFX-pythia8/MINIAODSIM/94X_mc2017_realistic_v10-v1/00000/005DC030-D3F4-E711-889A-02163E01A62D.root

=============================================

MessageLogger Summary

 type     category        sev    module        subroutine        count    total
 ---- -------------------- -- ---------------- ----------------  -----    -----
    1 XrdAdaptor           -w file_open                              1        1
    2 fileAction           -s file_close                             1        1
    3 fileAction           -s file_open                              2        2

 type    category    Examples: run/evt        run/evt          run/evt
 ---- -------------------- ---------------- ---------------- ----------------
    1 XrdAdaptor           pre-events                        
    2 fileAction           PostGlobalEndRun                  
    3 fileAction           pre-events       pre-events       

Severity    # Occurrences   Total Occurrences
--------    -------------   -----------------
Warning                 1                   1
System                  3                   3

dropped waiting message count 0

The output, ggtree_mc.root is the Ntuple that you can analyze.

Analyzing the Ntuples

One way to access the information contained in the ntuple directly is through the command line using ROOT interactively. Try these commands and draw the electron pt and see the effect of adding some cuts.

$ root -l ggtree_mc.root
root [1] .ls
root [2] ggNtuplizer->cd()
root [3] EventTree->Print()
root [4] EventTree->Draw("elePt")
root [5] EventTree->Draw("elePt","elePt>20") # add a cut at 20 GeV

You can apply a mask cut and select events with a tighter electron ID by doing this:

root [6] EventTree->Draw("elePt","(eleIDbit & 8)== 8")

You can learn a little more about eleIDbit and do more exercises in this link FNALHATS2017EGM). To understand more about the effects of the different levels of cuts, (Loose, Tight, Medium), let's superimpose the plots of the variable =sigmaIetaIetaFull5x5".

EventTree->SetLineColor(kBlack)
EventTree->Draw("eleSigmaIEtaIEtaFull5x5","(eleIDbit & 2)== 2")
EventTree->SetLineColor(kRed)
EventTree->Draw("eleSigmaIEtaIEtaFull5x5","(eleIDbit & 4)== 4","same")
EventTree->SetLineColor(kBlue)
EventTree->Draw("eleSigmaIEtaIEtaFull5x5","(eleIDbit & 8)== 8","same")

We can also use a simple plotting script to make the plots we did above.

https://raw.githubusercontent.com/uzzielperez/CMSSW-workbook-practice/master/ElectronAnalysis/simple_plotter.py

Using the simple plotter, we can reproduce and see the effect of putting a cut at 20 GeV. Making a few modifications with the script, we can make comparisons on the effect of different levels of cuts on different variables. In the script below, you can edit the variable name, "var".

wget https://raw.githubusercontent.com/uzzielperez/CMSSW-workbook-practice/master/ElectronAnalysis/EleIDSame.py

From the plots, what can you say about "fake" electrons in the high pT region? How does the "sigmaIetaIeta" distribution compare for electrons and jets?

Z->ee Exercise

The Z->ee channel is an important final state that is often used for calibration and systematic uncertainty measurements. In this exercise, we will be reconstructing the Z boson. We expect a resonance peak at around 91.1 GeV.

To do this, make sure you are in the CMSSW environment created for this page and copy a simple plotting script for this purpose.

cmsenv
voms-proxy-init -voms cms --valid 192:00
cd $CMSSW_BASE/src/ggAnalysis/ggNtuplizer/test
xrdcp root://cmseos.fnal.gov//store/user/drberry/HATS/makePlots.py . 
# or download a copy from Github 
wget 

The input file would be the Ntuple ggtree_mc.root which we made in the previous section. (Note: If you want to change the input file you can do python makePlots.py -I <filename.root>. To check command line options do, python makePlots.py -h). Running this script will output plots.root which has 3 the histograms for the pT, sigmaIetaIeta and the mass. Have a look at the mass distribution in particular and confirm that it peaks at around the expected mass of the Z.

Now, let's try to loosen and tighten the selection and correspondingly change the output names to plots_loose.root and plots_tight.root. This time, we'll be running on background MC.

python makePlots.py -o plots_loose.root -i root://cmsxrootd.fnal.gov//store/user/drberry/HATS/ggntuples/DYJetsToLL_M-50_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8.root root://cmsxrootd.fnal.gov//store/user/drberry/HATS/ggntuples/QCD_Pt-40toInf_DoubleEMEnriched_MGG-80toInf_TuneCUETP8M1_13TeV_Pythia8.root root://cmsxrootd.fnal.gov//store/user/drberry/HATS/ggntuples/QCD_Pt-40toInf_DoubleEMEnriched_MGG-80toInf_TuneCUETP8M1_13TeV_Pythia8.root

Change to the tight selection:

(event.eleIDbit[i]&8)!=8:

python makePlots.py -o plots_tight.root -i root://cmsxrootd.fnal.gov//store/user/drberry/HATS/ggntuples/DYJetsToLL_M-50_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8.root root://cmsxrootd.fnal.gov//store/user/drberry/HATS/ggntuples/QCD_Pt-40toInf_DoubleEMEnriched_MGG-80toInf_TuneCUETP8M1_13TeV_Pythia8.root root://cmsxrootd.fnal.gov//store/user/drberry/HATS/ggntuples/QCD_Pt-40toInf_DoubleEMEnriched_MGG-80toInf_TuneCUETP8M1_13TeV_Pythia8.root

To make a quick comparison of the effects of the loosening and the tightening of the selection, you can use a simple plotter that overlays the plots for several root files.

wget https://raw.githubusercontent.com/uzzielperez/CMSSW-workbook-practice/master/ElectronAnalysis/PltHistsSame.py

If we look at the output plot, much of the output background has been cut off. Let's start experimenting with different cuts and see how it affects the Z peak.

Note: If you have trouble accessing the files you can contact the most recent editor or try getting some practice on retrieving similar datasets from DAS and Ntuplize them to make a similar analysis above.

To make a quick comparison of the effects of the loosening and the tightening of the selection, you can use a simple plotter that overlays the plots for several root files.

wget https://raw.githubusercontent.com/uzzielperez/CMSSW-workbook-practice/master/ElectronAnalysis/PltHistsSame.py

If we look at the output plot, much of the output background has been cut off. Let's start experimenting with different cuts and see how it affects the Z peak. Here are some questions you can ask yourself:

• How does the mass peak look like for "barrel-only" and "endcap-only" electrons? For the barrel(endcap), the electron should fall in the abs(eleEta) < 1.4 (1.566 < eleEta && eleEta < 2.5) region.
• What is the effect of raising the R9 requirement? (R9 is the energy sum of 3x3 crystals in a supercluster divided by the raw energy within that cluster. The 3x3 crystals are centered on the most energetic crystal.)?
• Try other cuts on other variables like eleSigmaIEtaIEtaFull5x5, eleHoverE, elePFChIso, eleIDMVA.

You can also try comparing data with Montecarlo. You can try to list all the samples for this category

eosls /store/user/lpcsusystealth/DATA/ggSKIMS/ 

Let's run over the DoubleEG Run2016H sample. It has around 200M entries so we can choose to run over only 100000 of them.

python makePlots.py -o plots_data.root -m 100000 -i root://cmseos.fnal.gov//store/user/lpcsusystealth/DATA/ggSKIMS/DoubleEG_Run2016H_SepRereco_HLTDiPho3018M90_SKIM.root

We can vary fit parameters in the fitter code below in order to fit the data. This code is configured to run on MC so make sure you change the argument on the fitter function first. Try to see the difference between the regular and regressed energy.

xrdcp root://cmseos.fnal.gov//store/user/drberry/HATS/fitter.C .
root fitter.C

Tag and Probe Efficiency Measurement

The tag and probe method is a data-driven approach to measuring selection ID efficiencies. It exploits resonances that decay into two identical final states, like the JPsi or Z (to dilepton), to select a certain (tag) particles and probe the efficiency of a particular selection criterion being applied to them. The "tag" are often golden objects that need to pass a tight selection, has a very low fake rate, while the probe has a selection criteria that could be very loose. Resonances are used by reconstructing them in pairs with one final state being a "tag" (passing tight ID) while the other passes a loose ID is a probe. Lineshapes of (tag+"passing probes") and (tag+"passing probes") are then created and are separately fit with a signal + background model. The efficiency is then computed from the ratio of the number of "passing probes" and the total (passing + failing probes). These can be done in bins of different probe variables (e.g. pT, η ...).

Setup and PileUp

It is recommended that this part of the workbook be done at lxplus since many files that are needed here are stored at the CERN EOS. However, there's no reason that these cannot be run at FNAL. You can choose to use your cmslpc account instead if you do not have enough space in your lxplus account.

cmsrel CMSSW_9_4_0
cd CMSSW_9_4_0/src
cmsenv

Next we need to clone some GitHub code and update our TnPTreeproducer code. The parts below takes a while but we can save some time with the git cos-init with the following commands:

export CMSSW_GIT_REFERENCE=/cvmfs/cms.cern.ch/cmssw.git.daily #setenv CMSSW_GIT_REFERENCE /cvmfs/cms.cern.ch/cmssw.git.daily
git cms-init
git ls-remote --tags my-cmssw | cut -f2 | sed 's~refs/tags/~~' | xargs -n 1000 -P 1 git push my-cmssw --delete
git cms-merge-topic lsoffi:CMSSW_9_4_0_pre3_TnP
git cms-merge-topic guitargeek:ElectronID_MVA2017_940pre3
scram b -j8

# Add the area containing the MVA weights (from cms-data, to appear in “external”).
# Note: the “external” area appears after “scram build” is run at least once, as above
#
cd $CMSSW_BASE/external
# below, you may have a different architecture, this is just one example from lxplus
cd slc6_amd64_gcc630/
git clone https://github.com/lsoffi/RecoEgamma-PhotonIdentification.git data/RecoEgamma/PhotonIdentification/data
cd data/RecoEgamma/PhotonIdentification/data
git checkout CMSSW_9_4_0_pre3_TnP
cd $CMSSW_BASE/external
cd slc6_amd64_gcc630/
git clone https://github.com/lsoffi/RecoEgamma-ElectronIdentification.git data/RecoEgamma/ElectronIdentification/data
cd data/RecoEgamma/ElectronIdentification/data
git checkout CMSSW_9_4_0_pre3_TnP
# Go back to the src/
cd $CMSSW_BASE/src

cd $CMSSW_BASE/src
git clone -b CMSSW_9_4_X https://github.com/cms-analysis/EgammaAnalysis-TnPTreeProducer.git EgammaAnalysis/TnPTreeProducer

scram b -j4

We then need to retrieve the 2017 golden and luminosity JSON files from the CMS DQM area.

wget https://cms-service-dqm.web.cern.ch/cms-service-dqm/CAF/certification/Collisions17/13TeV/Final/Cert_294927-306462_13TeV_PromptReco_Collisions17_JSON.txt
wget https://cms-service-dqm.web.cern.ch/cms-service-dqm/CAF/certification/Collisions17/13TeV/PileUp/pileup_latest.txt

Using those two files we can now generate the 2017 data pileup distribution using a pileup calculator:

pileupCalc.py -i Cert_294927-306462_13TeV_PromptReco_Collisions17_JSON.txt --inputLumiJSON pileup_latest.txt --calcMode true --minBiasXsec 69200 --maxPileupBin 100 --numPileupBins 100 pileup_2017_41fb.root

Let's dump the values of the data pileup distribution to the std output.

cp ~charaf/nobackup/EgammaHATS2018/dumpPileup.py PhysicsTools/TagAndProbe/test/utilities/
python PhysicsTools/TagAndProbe/test/utilities/dumpPileup.py  pileup_2017_41fb.root

Replace the values for the key "2017_DATA_xSec69.2mb_94X_17Jan" in the "data_pu_distribs" dictionary.

data_pu_distribs = {.....
"2017_DATA_xSec69.2mb_94X_17Jan":[2.6e+05,1.08e+06,2.09e+06,3.69e+06,....

data_pu_distribs = {.....
"2017_DATA_xSec69.2mb_94X_17Jan":[2.6e+05,1.08e+06,2.09e+06,3.69e+06,....

   "2017_DATA_xSec69.2mb_94X_17Jan":[2.59e+05,1.08e+06,2.01e+06,3.78e+06,...,1.34,0.765,0.431]

Now we can produce the T&P trees!

Preparing the T&P dataset

To produce a T&P trees, you can do the following.

cd EgammaAnalysis/TnPTreeProducer/
cmsenv
cmsRun python/TnPTreeProducer_cfg.py doEleID=True isMC=False maxEvents=5000

Open the root file it produces which is called "TnPTree_data.root" and inspect a few distributions like = el_5x5_sieie, el_abseta, el_sc_eta, el_hoe=. The number of events in this file might not be enough for a "good" set of distributions so you can rerun this setting maxEvents to a higher number (~10000). Or you can retrieve a prepared T&P Tree by doing:

cd ~/cuperez/nobackup/CMSSWWorkbook/TnPTree_data_workbook.root .

Now we can run the fitter code to get the information we need from the tree.

Measuring Efficiencies by fitting to the dataset

We still need to download the tag and probe fitter code from GitHub.

cd $CMSSW_BASE/src
git clone -b egm_tnp_Moriond18_v3.0 git@github.com:michelif/egm_tnp_analysis.git
cd egm_tnp_analysis/
make

We'll be modifying some of the python files so it's good to keep a backup copy:

cp etc/config/settings_ele.py etc/config/settings_ele_workbook.py 
cp etc/config/tnpSampleDef.py etc/config/tnpSampleDef_workbook.py 
cp etc/config/tnpEGM_fitter.py etc/config/tnpEGM_fitter_workbook.py

If you are at FNAL LPC, you need to replace eos/cms by root://eoscms.cern.ch/. If at CERN, put / before eos/cms. The need to use xrootd at LPC is a lot slower so it is recommended that you work at lxplus.

To run the fitter code, try the following commands and open egammaEffi.txt to view the results. Note that --mcSig --altSig" "--altSig" may take at least an hour.

python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X --checkBins
python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X --createBins
python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X --createHists
python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X --doFit
python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X --doFit --mcSig --altSig
python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X --doFit --altSig
python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X --doFit --altBkg
python tnpEGM_fitter_workbook.py etc/config/settings_ele_workbook.py --flag passingTight94X  --sumUp

NanoAOD

It's good to familiarize oneself with the recently developed NanoAOD data tier. You can learn more about it in the WorkBookNanoAOD. The goal for this new data tier is to keep the event size below 2kb/event so that ~50% of analyses can have their ntuples centrally produced. We can run through a quick example and make a quick analysis with them.

cd $CMSSW_BASE/src 
cmsenv 
git cms-init # not really needed except if you want other cmssw stuff
git clone https://github.com/cms-nanoAOD/nanoAOD-tools.git PhysicsTools/NanoAODTools
scram b
voms-proxy-init -voms cms 

cd PhysicsTools/NanoAODTools/python/postprocessing/examples/
python exampleAnalysis.py

Feel free to look at the code and how you could make selections (e.g. choose events with at least two electrons) as you loop over the events. You could try finding another interesting dataset in DAS and look at what's stored in the root files.

...

Miscellaneous Links

Workshops and Tutorials:

• CMS Data Analysis School LPC 2018: E/gamma Objects - https://twiki.cern.ch/twiki/bin/viewauth/CMS/SWGuideCMSDataAnalysisSchoolLPC2018egamma
• CMS Hands on Tutorial LPC 2017: Photons and electrons - https://twiki.cern.ch/twiki/bin/view/CMS/FNALHATS2017EGM#Create_Ntuples
• Electron/Photon HATS@LPC (Session I) : June 28, 2018 - https://indico.cern.ch/event/727032/
• Electron/Photon HATS@LPC (Session II): June 29, 2018 - https://indico.cern.ch/event/727034/

Miscellaneous:

• ggNtuplizer - https://github.com/cmkuo/ggAnalysis/tree/94X
• Implementation details for electron and photon ID: https://twiki.cern.ch/twiki/bin/view/CMS/EgammaIDRecipesRun2
• CMS Data Aggregation System - https://cmsweb.cern.ch/das/
• Electron Tag and Probe - https://twiki.cern.ch/twiki/bin/view/CMSPublic/ElectronTagAndProbe

Review status

Reviewer/Editor and Date (copy from screen)	Comments
UzzielPerez - 15-Aug-2018	Update for Run 2, CMSSW 94x
MarcoPieri - 24-Jan-2010
DavidChamont - End November 2009	Upgrade for 3XX.
MatteoSani - 02 Mar 2009
ChrisSeez - 22 Feb 2008
Main.Aresh - 27 Feb 2008	Changes in verbatim elements (in the current page and also in https://twiki.cern.ch/twiki/bin/view/CMS/WorkBookElectronAnalysis) because of some lines too long for printable version
JennyWilliams - 08 Oct 2007	recreate contents table
Main.palmale - 05 Dec 2006	Review and edit

Responsible:Main.DavidChamont