gromacs常见错误

Thevast巨大的majorityoferrormessagesgeneratedbyGROMACSaredescriptive解释的,informingtheuserwheretheexacterrorlies.Someerrorsthatarisearenotedbelow,alongwithmoredetailsonwhattheissueisandhowtosolveit.1.1.General1.1.1.Cannotallocatememory2.2.pdb2gmx1.2.1.Residue'XXX'notfoundinresiduetopologydatabase2.2.2.Longbondsand/ormissingatoms3.2.3.Chainidentifier'X'wasusedintwonon-sequentialblocks4.2.4.WARNING:atomXismissinginresidueXXXYinthepdbfile5.2.5.AtomXinresidueYYYnotfoundinrtpentry3.3.grompp1.3.1.Foundaseconddefaultsdirectivefile2.3.2.Invalidorderfordirectivedefaults3.3.3.Systemhasnon-zerototalcharge4.3.4.Incorrectnumberofparameters5.3.5.Numberofcoordinatesincoordinatefiledoesnotmatchtopology6.3.6.Fatalerror:NosuchmoleculetypeXXX7.3.7.T-CouplinggroupXXXhasfewerthan10%oftheatoms8.3.8.Thecut-offlengthislongerthanhalftheshortestboxvectororlongerthanthesmallestboxdiagonalelement.Increasetheboxsizeordecreaserlist9.3.9.Unknownleft-handXXXXinparameterfile10.3.10.Atomindex(1)inbondsoutofbounds4.4.mdrun1.4.1.Stepsizetoosmall,ornochangeinenergy.Convergedtomachineprecision,butnottotherequestedprecision2.4.2.LINCS/SETTLE/SHAKEwarnings3.4.3.1-4interactionnotwithincut-off4.4.4.Simulationrunningbutnooutput5.4.5.CannotdoConjugateGradientswithconstraints6.4.6.Pressurescalingmorethan1%7.4.7.RangeCheckingerror8.4.8.XparticlescommunicatedtoPMEnodeYaremorethanacelllengthoutofthedomaindecompositioncelloftheirchargegroup9.4.9.ThereisnodomaindecompositionfornnodesthatiscompatiblewiththegivenboxandaminimumcellsizeofxnmGeneral CannotallocatememoryTheexecutedscripthasattemptedtoassignmemorytobeusedinthecalculation,butisunabletoduetoinsufficientmemory.Possiblesolutionsare:·installmorememoryinthecomputer.·useacomputerwithmorememory.·reducethescopeofthenumberofatomsselectedforanalysis.·reducethelengthoftrajectoryfilebeingprocessed.·insomecasesconfusionbetweenÅngströmandnmmayleadtouserswantingtogenerateawaterboxthatis103timeslargerthanwhattheythinkitis(e.g.genbox).Theusershouldbearinmindthatthecostintimeand/ormemoryforvariousactivitieswillscalewiththenumberofatoms/groups/residuesNorthesimulationlengthTasorderN,NlogN,orN2(ormaybeworse!)andthesameforT,dependingonthetypeofactivity.Ifittakesalongtime,haveathinkaboutwhatyouaredoing,andtheunderlyingalgorithm(seethemanual,manpage,orusethe-hflagfortheutility),andseeifthere'ssomethingsensibleyoucandothathasbetterscalingproperties. pdb2gmx Residue'XXX'notfoundinresiduetopology残基不存在databaseThismeansthattheforcefield{youhaveselectedwhilerunningpdb2gmx}doesnothaveanentryintheresiduedatabaseforXXX.Theresiduedatabaseentryisnecessarybothforstand-alonemolecules(e.g.formaldehyde甲醛)orapeptide(standardornon-standard).Thisentrydefinestheatomtypes,connectivity,bondedandnon-bondedinteractiontypesfortheresidueandisnecessarytousepdb2gmxtobuilda.topfile.Aresiduedatabaseentrymaybemissingsimplybecausethedatabasedoesnotcontaintheresidueatall,orbecausethenameisdifferent.Fornewusers,thiserrorappearsbecausetheyarerunningpdb2gmxblindly盲目的onaPDBfiletheyhavewithoutconsiderationofthecontentsofthefile.Aforcefieldisnotsomethingthatismagical,itcanonlydealwithmoleculesorresidues(buildingblocks)thatareprovidedintheresiduedatabaseorincludedotherwise.Ifyouwanttousepdb2gmxtoautomaticallygenerateyourtopology,youhavetoensurethattheappropriate.rtpentryispresentwithinthedesiredforcefieldandhasthesamenameasthebuildingblockyouaretryingtouse.Ifyoucallyourmolecule"HIS,"thenpdb2gmxwillnotmagicallybuildarandommolecule;itwilltrytobuildhistidine,basedonthe[HIS]entryinthe.rtpfile,soitwilllookfortheexactatomicentriesforhistidine,nomorenoless.Ifyouwantatopologyforanarbitrarymolecule,youcannotusepdb2gmx(unlessyoubuildthe.rtpentryyourself).Youwillhavetobuilditbyhand,oruseanotherprogram(suchasx2toporoneofthescriptscontributedbyusers)tobuildthe.topfile.Ifthereisnotanentryforthisresidueinthedatabase,thentheoptionsforobtainingtheforcefieldparametersare:·seeifthereisadifferentnamebeingusedfortheresidueintheresiduedatabaseandrenameasappropriate,·parameterizetheresidue/moleculeyourself,·findatopologyfileforthemolecule,convertittoan.itpfileandincludeitinyour.topfile,·useanotherforcefieldwhichhasparametersavailableforthis,·searchtheprimaryliteratureforpublicationsforparametersfortheresiduethatareconsistentwiththeforcefieldthatisbeingused. Longbondsand/ormissingatomsThereareprobablyatomsmissingearlierinthe.pdbfilewhichmakespdb2gmxgocrazy.Checkthescreenoutputofpdb2gmx,asitwilltellyouwhichoneismissing.Thenaddtheatomsinyourpdbfile,energyminimizationwillputthemintherightplace,orfixthesidechainwithe.g.theWhatIFprogram. Chainidentifier'X'wasusedintwonon-sequential非次序的blocksThismeansthatwithinthecoordinatefilefedtopdb2gmx,theXchainhasbeensplit断裂,possiblybytheincorrectinsertionofonemoleculewithinanother.Thesolutionissimple:movetheinsertedmoleculetoalocationwithinthefilesothatitisnotsplittinganothermolecule.Thismessagemayalsomeanthatthesamechainidentifierhasbeenusedfortwoseparatechains.Inthatcase,renamethesecondchaintoauniqueidentifier. WARNING:atomXismissinginresidueXXXYinthepdbfileRelatedtothelongbonds/missingatomserrorabove,thiserrorisusuallyquiteobviousinitsmeaning.Thatis,pdb2gmxexpectscertainatomswithinthegivenresidue,basedontheentriesintheforcefield.rtpfile.Thereareseveralcasestowhichthiserrorapplies:·Missinghydrogenatoms;theerrormessagemaybesuggestingthatanentryinthe.hdbfileismissing. Morelikely,thenomenclature名称ofyourhydrogenatomssimplydoesnotmatchwhatisexpectedbythe.rtpentry. Inthiscase,use-ignhtoallowpdb2gmxtoaddthecorrecthydrogensforyou,orre-nametheproblematic问题的atoms.·Aterminalresidue(usuallytheN-terminus)ismissingHatoms;thisusuallysuggeststhattheproper-teroptionhasnotbeensuppliedorchosenproperly.InthecaseoftheAMBERforcefields,nomenclatureistypicallytheproblem.N-terminalandC-terminalresiduesmustbeprefixedbyNandC,respectively. Forexample,anN-terminalalanine丙氨酸shouldnotbelistedinthe.pdbfileasALA,butratherNALA,asspecifiedintheffamberinstructions.·Atomsaresimplymissinginthestructurefileprovidedtopdb2gmx;lookforREMARK465andREMARK470entriesinthe.pdbfile.Theseatomswillhavetobemodeledinusingexternalsoftware.ThereisnoGROMACStooltore-constructincompletemodels.Contrarytowhattheerrormessagesays,theuseoftheoption-missingisalmostalwaysinappropriate. The-missingoptionshouldonlybeusedtogeneratespecializedtopologiesforaminoacid-likemoleculestotakeadvantageof.rtpentries. Ifyoufindyourselfusing-missinginordertogenerateatopologyforaproteinornucleicacid,don't;thetopologyproducedislikelyphysicallyunrealistic. AtomXinresidueYYYnotfoundinrtpentryIfyouareattemptingtoassembleatopologyusingpdb2gmx,theatomnamesareexpectedtomatchthosefoundinthe.rtpfile thatdefinethebuildingblock(s)inyourstructure.  Ifyougetthiserror,simplyre-nametheatomsinyourcoordinatefileappropriately. grompp FoundaseconddefaultsdirectivefileThisiscausedbythe[defaults]directiveappearingmorethanonceinthetopologyorforcefieldfilesforthesystem-itcanonlyappearonce.Atypicalcauseofthisisaseconddefaultsbeingsetinanincludedtopologyfile,.itp,thathasbeensourcedfromsomewhereelse.Forspecificationsonhowthetopologyfileswork,seeGROMACSManual,Section5.6.[defaults];nbfunccomb-rulegen-pairsfudgeLJfudgeQQ11no1.01.0Onesolutionistosimplycommentout(ordelete)thelinesofcodeoutinthefilewhereitisincludedforthesecondtimei.e.,;[defaults];nbfunccomb-rulegen-pairsfudgeLJfudgeQQ;11no1.01.0Abetterapproachtofindingasolutionistore-thinkwhatyouaredoing.The[defaults]directiveshouldonlybeappearingatthetopofyour.topfilewhereyouchoosetheforcefield.Ifyouaretryingtomixtwoforcefields,thenyouareaskingfortrouble.Ifamolecule.itpfiletriestochooseaforcefield,thenwhoeverproduceditisaskingfortrouble. Invalid无效的orderfordirectivedefaultsThisisaresultofdefaultsbeingsetinthetopologyorforcefieldfilesintheinappropriatelocation;the[defaults]sectioncanonlyappearonceandmustbethefirstdirectiveinthetopology. The[defaults]directiveistypicallypresentintheforcefieldfile(ffgmx.itp,ffoplsaa.itp,etc),andisaddedtothetopologywhenyou#includethisfileinthesystemtopology.The"invalidorder"errorcanalsopertainto适合anyoftheothertopologydirectives. Inthesecases,theerrorisusuallyaresultofsimplyinappropriatelyre-organizingthetopology(hence"outoforder"or"outofsequence,"astheerrorimplies). Acommonproblemisplacingpositionrestraintfilesformultiplemoleculesoutoforder. Recallthatapositionrestraintfilecanonlybelongtothe[moleculetype]blockthatcontainsit. Forexample:WRONG:#include"topol_A.itp"#include"topol_B.itp"#include"ligand.itp"#ifdefPOSRES#include"posre_A.itp"#include"posre_B.itp"#include"ligand_posre.itp"#endifRIGHT:#include"topol_A.itp"#ifdefPOSRES#include"posre_A.itp"#endif#include"topol_B.itp"#ifdefPOSRES#include"posre_B.itp"#endif#include"ligand.itp"#ifdefPOSRES#include"ligand_posre.itp"#endif Systemhasnon-zerototalchargeNotifiesyouthatcounter-ions反离子mayberequiredforthesystemtoneutralizethechargeortheremaybeproblemswiththetopology.Ifthechargeisanon-integer,thenthisindicatesthatthereisaproblemwiththetopology.Ifpdb2gmxhasbeenused,thenlookattherighthandcommentcolumnoftheatomlisting,whichliststhecumulative累计charge.Thisshouldbeanintegeraftereveryresidue(and/orchargegroupwhereapplicable).Thiswillassistinfindingtheresiduewherethingsstartdepartingfromintegervalues.Alsocheckthecappinggroupsthathavebeenused.Ifthechargeisalreadyclosetoaninteger,thenthedifferenceiscausedbyroundingerrorsandnotamajorproblem.NoteforPMEusers:ItispossibletouseauniformneutralizingbackgroundchargeinPMEtocompensateforasystemwithanetbackgroundcharge.Thereisprobablynothingwrongwiththisinprinciple,becausetheuniformchargewillnotperturb不安thedynamics.Nevertheless,itisstandardpracticetoactuallyaddcounter-ionstomakethesystemnetneutral. IncorrectnumberofparametersLookatthetopologyfileforthesystem.You'venotgivenenoughparametersforoneofthebondeddefinitions.Sometimesthisalsooccursifyou'vemangled损坏theIncludeFileMechanismorthetopologyfileformat(see:GROMACSManualChapter5)whenyoueditedthefile. NumberofcoordinatesincoordinatefiledoesnotmatchtopologyThisispointingoutthat,basedontheinformationprovidedinthetopologyfile,.top,thetotalnumberofatomsorparticleswithinthesystemdoesnotmatchexactlywithwhatisprovidedwithinthecoordinatefile,oftena.groora.pdb.Themostcommonreasonforthisissimplythattheuserhasfailedtoupdatethetopologyfileaftersolvatingoraddingadditionalmoleculestothesystem,ormadeatypographicalerrorinthenumberofoneofthemoleculeswithinthesystem.Ensurethattheendofthetopologyfilebeingusedcontainssomethinglikethefollowing,thatmatchesexactlywithwhatiswithinthecoordinatefilebeingused,intermsofbothnumbersandorderofthemolecules:[molecules];Compound#molProtein1SOL10189NA+10Inacasewhengromppcan'tfindanyanyatomsinthetopologyfileatall(numberofcoordinatesincoordinatefile(conf.gro,648)doesnotmatchtopology(topol.top,0))andthaterrorisprecededbywarningslike:calling/lib/cpp...sh:/lib/cpp:Nosuchfileordirectorycppexitcode:32512Triedtoexecute:'/lib/cpp-I/usr/local/gromacs-...The'/lib/cpp'commandisdefinedinthe.mdpfilethenyoursystem'sCpreprocessor,cpp,isnotbeingfoundorruncorrectly.Onereasonmightalsobethatthecppvariableisnotproperlysetinthe.mdpfile. AsofGROMACSversion4.0,gromppcontainsitsownpreprocessor,sothiserrorshouldnotoccur.Thiserrorcanalsooccurwhenthe.mdpfilehasbeeneditedunderWindows,andyourcppisintolerantofthemismatchbetweenWindowsandUnixend-of-linecharacters.Ifitispossiblethatyouhavedonethis,tryrunningyour.mdpfilethroughthestandardLinuxdos2unixutility. Fatalerror:NosuchmoleculetypeXXXEachtypeofmoleculeinyour[molecules]sectionofyour.topfilemusthaveacorresponding[moleculetype]sectiondefinedpreviously,eitherinthe.topfileoranincluded.itpfile.SeeGROMACSManualsection5.6.1forthesyntax句法description.Your.topfiledoesn'thavesuchadefinitionfortheindicatedmolecule.Checkthecontentsoftherelevantfiles,andpayattentiontothestatusof#ifdef and/or#includestatements. T-CouplinggroupXXXhasfewerthan10%oftheatomsItispossibletospecifyseparatethermostats 恒温器(temperaturecouplinggroups)foreachandeverymoleculetypewithinasimulation.ThisisaparticularlybadpracticeemployedbymanynewuserstoMolecularDynamicsSimulations. Doingsoisabadidea,asyoucanintroduceerrorsandartifactsthatarehardtopredict.Inmostcasesitisbesttohaveallmoleculeswithinasinglegroup,usingsystem.Ifseparatecouplinggroupsarerequired,thenensurethattheyareof"sufficientsize"andcombinemoleculetypesthatappeartogetherwithinthesimulation.Forexample,foraproteininwaterwithcounter-ions,onewouldlikelywanttouseProteinandNon-Protein. Thecut-offlengthislongerthanhalftheshortestboxvectororlongerthanthesmallestboxdiagonalelement.IncreasetheboxsizeordecreaserlistThiserrorisgeneratedinthecasesasnotedwithinthemessage.Thedimensionsoftheboxaresuchthatanatomwillinteractwithitself(whenusingperiodicboundaryconditions),thusviolatingtheminimumimageconvention.Suchaneventistotallyunrealisticandwillintroducesomeseriousartefacts.Thesolutionisagainwhatisnotedwithinthemessage,eitherincreasethesizeofthesimulationboxsothatitisatanabsoluteminimumtwicethecut-offlengthinallthreedimensions(takecarehereifareusingpressurecoupling,astheboxdimensionswillchangeovertimeandiftheydecreaseevenslightly,youwillstillbeviolatingtheminimumimageconvention)ordecreasethecut-offlength(dependingontheforcefieldutilised,thismaynotbeanoption). Unknownleft-handXXXXinparameterfilegrompphasfoundanunknownterminthe.mdpfilefedtoit.YoushouldcheckthespellingofXXXXandlookfortypographical排字上的errors.BeawarethatquiteafewrunparameterschangedbetweenGROMACS3.xandGROMACS4.xandtheoutputfromgromppwillsometimesofferhelpfulcommentaryaboutthesesituations. Atomindex(1)inbondsoutofboundsThiskindoferrorlookslikeFatalerror:[filespc.itp,line32]Atomindex(1)inbondsoutofbounds(1-0).Thisprobablymeansthatyouhaveinsertedtopologysection"settles"inapartbelongingtoadifferentmoleculethanyouintendedto.inthatcasemovethe"settles"sectiontotherightmolecule.Thiserrorisfairlyself-explanatory.Youshouldlookatyour.topfileandcheckthatallofthe[molecules]sectionscontainallofthedatapertainingtothatmolecule,andnootherdata.Thatis,youcannot#includeanothermoleculetype(.itpfile)beforetheprevious[moleculetype]hasended. Consulttheexamplesinchapter5ofthemanualforinformationontherequiredorderingofthedifferent[sections].Payattentiontothecontentsofanyfilesyouhaveincludedwith#includedirectives.Thiserrorcanalsoariseifyouareusingawatermodelthatisnotenabledforusewithyourchosenforcefieldbydefault. Forexample,ifyouareattemptingtousetheSPCwatermodelwithanAMBERforcefield,youwillseethiserror. Thereasonisthat,inspc.itp,thereisno#ifdefstatementdefiningatomtypesforanyoftheAMBERforcefields. Youcaneitheraddthissectionyourself,oruseadifferentwatermodel. mdrun Stepsizetoosmall,ornochangeinenergy.Convergedtomachineprecision,butnottotherequestedprecisionThisisnotanerrorassuch.Itissimplyinformingyouthatduringtheenergyminimizationprocessitreachedthelimitpossibletominimizethestructurewithyourcurrentparameters.Itdoesnotmeanthatthesystemhasnotbeenminimizedfully,butinsomesituationsthatmaybethecase.Ifthesystemhasasignificantamountofwaterpresent,thenaEpotoftheorderof-105to-106istypicallyareasonablevalueforalmostallcases,e.g.startingaMolecularDynamicssimulationfromtheresultingstructure.Onlyforspecialpurposes,suchasnormalmodeanalysistypeofcalculations,itmayberequiredtominimizefurther.Furtherminimizationmaybeachievedbyusingadifferentenergyminimizationmethodorbymakinguseof doubleprecision-enabled GROMACS. LINCS/SETTLE/SHAKEwarningsSometimes,whenrunningdynamics,mdrunmaysuddenlystop(perhapsafterwritingseveralpdbfiles)afteraseriesofwarningsabouttheconstraintalgorithms(e.g.LINCS,SETTLEorSHAKE)arewrittentothelogfile.Thesealgorithmsoftenusedtoconstrainbondlengthsand/orangles.Whenasystemisblowingup(i.e.explodingduetodivergingforces),theconstraintsareusuallythefirstthingtofail.Thisdoesn'tnecessarilymeanyouneedtotroubleshoottheconstraintalgorithm.Usuallyitisasignofsomethingmorefundamentallywrong(physicallyunrealistic)withyoursystem.Perhapsyoudidn'tminimizewellenough,haveabadstartingstructure/stericclashes,areusingtoolargeatimestep,oraredoingparticleinsertioninfreeenergycalculationswithoutusingsoftcore.Thiscanalsobecausedbyasinglewatermoleculethatisisolatedfromtheotherwatermolecules,somewherewithinthesystem. 1-4interactionnotwithincut-offSomeofyouratomshavemovedsotwoatomsseparatedbythreebondsareseparatedbymorethanthecut-offdistance.ThisisBAD.Mostimportantly,donotincreaseyourcut-off!Thiserroractuallyindicatesthattheatomshaveverylargevelocities,whichusuallymeansthat(partof)yourmolecule(s)is(are)blowingup.IfyouareusingLINCSforconstraints,youprobablyalsoalreadygotanumberofLINCSwarnings.WhenusingSHAKEthiswillgiverisetoaSHAKEerror,whichhalts暂停yoursimulationbeforethe"1-4notwithincutoff"errorcanappear.Therecanbeanumberofreasonsforthelargevelocitiesinyoursystem.Ifithappensatthebeginningofthesimulation,yoursystemmightbenotequilibratedwellenough(e.g.itcontainssomebadcontacts).Trya(nother)roundofenergyminimizationtofixthis.Otherwiseyoumighthaveaveryhightemperature,and/oratimestep thatistoolarge.Experimentwiththeseparametersuntiltheerrorstopsoccurring.Ifthisdoesn'thelp,checkthevalidityoftheparametersinyourtopology! SimulationrunningbutnooutputNotanerrorassuch,butmdrunappearstobechewingupCPUtimebutnothingisbeingwrittentotheoutputfiles.Thereareanumberofreasonswhythismayoccur:·Yoursimulationmightsimplybe(very)slow,andsinceoutputisbuffered减轻,itcantakequitesometimeforoutputtoappearintherespectivefiles.Ifyouaretryingtofixsomeproblemsandyouwanttogetoutputasfastaspossible,youcansettheenvironmentvariableLOG_BUFSto0byusingsetenvLOG_BUFS0,thisdisablesoutputbuffering.UseunsetenvLOG_BUFStoturnbufferingonagain.·Somethingmightbegoingwronginyoursimulation,causinge.g.not-a-numbers(NAN)tobegenerated(thesearetheresultofe.g.divisionbyzero).SubsequentcalculationswithNAN'swillgeneratefloatingpointexceptionswhichsloweverythingdownbyordersofmagnitude.OnaSGIsystemthiswillusuallyresultinalargepercentageofCPUtimebeingdevotedto'system'(checkitwithosview,orforamulti-processormachinewithtopandosview).·Youmighthaveallnst*parameters(seeyour.mdpfile)setto0,thiswillsuppress阻止mostoutput.·Yourdiskmightbefull.Eventuallythiswillleadtomdruncrashing,butsinceoutputisbuffered,itmighttakeawhileformdruntorealizeitcan'twrite.·Youarerunninganexecutable可执行的compiledwithMPIsupport(e.g.LAM)anddidnotstarttheLAMdaemon无交互后台程序(lamboot).SeeLAMdocumentation. CannotdoConjugate结合GradientswithconstraintsThismeansyoucan'tdoenergyminimizationwiththeconjugategradientalgorithm共轭斜度算法ifyourtopologyhasconstraintsdefined-seehere. Pressurescalingmorethan1%Thiserrortendstobegeneratedwhenthesimulationboxbeginstooscillate摆动(duetolargepressuresand/orsmallcouplingconstants),thesystemstartstoresonate共振andthencrashes.Thiscanmeanthatthesystemisn'tequilibratedsufficientlybeforeusingpressurecoupling.Therefore,better/moreequilibrationmayfixtheissue.Itisrecommendedtoobservethesystemtrajectorypriorandduringthecrash.Thismayindicateifaparticularpartofthesystem/structureistheproblem.Insomecases,ifthesystemhasbeenequilibratedsufficiently,thiserrorcanmeanthatthepressurecouplingconstant,tau_p,istoosmall(particularlywhenusingtheBerendsenweakcouplingmethod).Increasingthatvaluewillslowdowntheresponsetopressurechangesandmaystoptheresonancefromoccuring.Thiserrorcanalsoappearwhenusingatimestepthatistoolarge,e.g.5fs,intheabsenceofconstraintsand/orvirtualsites. RangeCheckingerrorThisusuallymeansyoursimulationisblowingup.Probablyyouneedtodobetterenergyminimizationand/orequilibrationand/ortopologydesign. XparticlescommunicatedtoPMEn