rcsb.org
D3RGrandChallenge2StructureRefinementExperience/LearningsStephenK.Burley,M.D.,D.Phil.Director,RCSBProteinDataBankMember,RutgersCancerInstituteofNewJerseyFoundingDirector,InstituteforQuantitativeBiomedicineDirector,CenterforIntegrativeProteomicsResearchDistinguishedProfessor,ChemistryandChemicalBiologyD3RWebinarMeetingMarch27th2017
Outline
§ ProteinDataBank(RCSBPDB,rcsb.org)§ RCSBPDBGroupDepositionSystem§ D3RStructureRePinementProtocol§ Experience/Learnings§ Acknowledgements
1
ProteinDataBankArchive
§ PDB1stOpenAccessdigitaldataresourceinallofbiology
§ Founded1971with7X-raystructures
§ Today,singleglobalarchive
§ ManagedbywwPDBPartnership(RCSBPDB,PDBe,PDBj,BMRB)
2
Lysozyme!
Hemoglobin!
Ribonuclease!
Myoglobin!
Some of the very first structures in the PDB
PDBFactsandFigures
§ >128,000StructuresArchivedsince1971§ ~11,500NewStructuresDeposited/Year§ ~30,000DepositorsWorldwide§ >1MillionUniqueUsersWorldwide/Year§ >550MillionDataFilesDownloaded/Year§ ~1.5MillionDataFilesDownloaded/Day§ USPDBHeadquarteredatRutgers/UCSD§ ~70%ofGlobalPDBUsagethroughRutgers/UCSD
3
RCSBPDBGroupDeposiEonSystemhttps://deposit-group.rcsb.rutgers.edu/groupdeposit/ § SupportsBatchDeposition/ValidationofRelatedStructures§ OriginallydevelopedforD3RBlindChallenges§ Todate,>1300structuresfromRoche,MerckSerono,BMS,SGC,etc.,with10s-100sofstructures/group
4
OneDep DepUI
Batch deposition: LogIn, Upload,
Download, Validation, Communication
Issue PDB & Dep ids
Status DB
Batch processing: Scripts/Special
treatment + stand-alone OneDep modules
OneDep Biocuration
DB Load
Release Module
wwPDB OneDep System
RCSB PDB GroupDep System
D3RProtocol
§ D3RcompletesrePinements
§ GroupDepàPDB
§ PDBValidatesallrePinedstructures
§ ContributorsandD3Rco-authorPinalPDBdeposition
5
DataFilesfromContributors
D3RQACheck
Fail
Pass
ValidationOK?
D3RRePinement
PDBDeposition(GroupDep)
wwPDBValidationReport
FurtherProtein/LigandRePinement
D3RQualityReport
PDBHold/Release
D3RChallengeParticipants
Yes
No
FinalContributorReview
D3RStructureRefinementProtocol
6
SFFiles(AnyFormat)
ModelFiles(CIF/PDB)
LigandRestraints(CIF)
Re+inementRefmac,
Phenix,Buster(various
parametersoptimized)
EDMapReview
(RSR,RSCC,RSZDforbackbone,sidechains)
wwPDBValidation(EDmaps,geometry,sequence,ligands)
Visualchecksofproblemareas
ManualcorrectionusingCOOT
RocheFXRDataSetCharacterisEcs
§ 36Co-crystal/1ApoStructures• ResolutionLimits:2.6-1.7Å• AsymmetricUnitMWs:26-110kDa• 3PointGroups/5SpaceGroups
§ LigandMWs:219-564Da§ ThreeLigands:Bratoms§ FourLigands:AlternativeConformations§ AllLigands:BuriedinBindingPocket
7
ProteinStructureQualityImproved
8
InitialRange FinalRange AverageChange
Rfree(%) 23.0-35.0 22.0-31.0 -1.0(better)Rwork(%) 17.0-30.0 18.0-25.0 <1.0(nochange)Rfree–Rwork(%) 4.0-11.0 3.0-10.0 -1.0(better)ClashScore 4.38-40.41 0.51-6.90 -15.16(better)RamachandranOutliers(%)
0.0-2.2 0.0-1.3 -0.3(better)
SidechainRotamerOutliers(%)
0.5-24.5 0.5-9.9 -7.9(better)
%RSRZOutliers 0.0-29.5 0.0-22.1 -2.2(better)BondsRMSZ 0.3-0.7 0.4-0.9 +0.2(worse)AnglesRMSZ 0.5-0.8 0.5-0.9 +0.1(worse)OverallQuality(%ileofarchive,seeShaoetal.2017)
0%-70% 5%-80% +25%(better)
LigandStructureQualityImproved
9
Initial Final AverageChange
RSR(%) 8.0–21.0 7.0–19.0 -1.0(better)
RSCC(%) 82.0–98.0 79.0–98.0 +1.0(better)
BondLengthsOutliers
0–17 0-10 -3(better)
BondAnglesOutliers
3–24 0–7 -7(better)
Clashes 0-8 0-4 -1(better)
BondsRMSZ 0.8–2.9 0.5-2.9 -0.5(better)
AnglesRMSZ 1.1-3.1 0.5-2.0 -0.7(better)
RSZD+(absolute) 0.0-2.8 0.0-3.4 +0.2(balanced)
RSZD-(absolute) 0.0-4.2 0.0-3.3 -0.2(balanced)
LigandGeometricalRestraints
10
§ Generatedligandrestraintswith• AM1andRM1QMoptimizers
• +/-Mogulvalues• Variedplanaritysettings
§ BestresultswithAM1+Mogulandtighterplanarity,andmanualsettingofsomerestraintsasneeded
Roche1ymrc
Experience/Learnings:ParEalDisorder
11
Partialdisorderalternativeconformers(1dqzc,2.0Å)ThiophenealternativeconformationsCyclohexanedisorderedBlue:2|Fo|-|Fc|/1.0σ
Experience/Learnings:OverallDisorder
12
Entireligandintwoorientations(1txca,1.9Å)Blue:2|Fo|-|Fc|/1.0σ
Experience/Learnings:LossofBr
13
LigandwithBr(1xxar,1.8Å)Br-CbondoftenbrokenbyX-raybeamCrystalcontainstwoLigandspecies(+/-Br)BrModeledwithreducedOccupancyBlue:2|Fo|-|Fc|/1.5σRed:-ve|Fo|-|Fc|/3.0σGreen:+ve|Fo|-|Fc|/3.0σ
BrOccup=1.0 BrOccup=0.5
Experience/Learnings:AmbiguiEes
14
EDMapAmbiguity(fromdifferentstudy)Pseudo-symmetricligandsrequireinspectionofbindingenvironmentWrongOrientation!Blue:2|Fo|-|Fc|/1.0σ
Acknowledgements
§ HuanwangYang,ChenghuaShao,ZukangFeng,JohnWestbrook,JasmineYoung,EzraPeisach
§ SymonGathiaka,ShuaiLiu,MikeGilson,RommieAmmaro
15The RCSB PDB is funded by a grant from the National Science Foundation, the National Institutes of Health, and the US Department of Energy.
