optim algorithm function deimmun therapeut protein character space interatom distanc distribut function consist solut scatter data abstract scatter neutron rai molecul solut offer altern approach studi wide rang macromolecular structur solut state crystal paper studi part problem elucid dimension structur solut scatter data determin distribut interatom distanc problem ill condit singl observ diffract pattern consist distanc distribut function due ill condit risk overfit observ scatter data propos approach avoid problem accept valid multipl altern curv seek singl show linear constraint ensur comput consist experiment data constraint enforc smooth curv ensur curv probabl distribut experiment error constraint precis describ space consist curv polytop histogram valu fourier coeffici descript sampl space potenti altern curv descript develop linear program approach sampl space consist realist curv test experiment simul scatter data approach effici gener ensembl protein fragment swap method asymmetr abstract paper present approach site direct recombin swap combin select discontigu fragment sourc protein place fragment target protein asymmetr differenti sourc target select swap discontigu fragment method focus experiment effort restrict portion sequenc space construct hybrid properti object experi sourc target structur homolog local method support swap fragment function import region sourc target scaffold human exogen therapeut protein protein fragment swap plan defin residu posit boundari fragment swap assess averag potenti score result hybrid librari singleton pairwis term evalu import fit swap residu prove np hard choos optim set fragment potenti score develop integ program approach call swagmer work practic demonstr effect method type swap problem select recombin beta lactamas activ swap glutathion transferas show select recombin approach gener plan term result potenti score tradit site direct recombin approach show case optim experi significantli result stochast method protein design sampl undirect graphic model residu constraint protein engin seek produc amino acid sequenc desir characterist structur function difficult problem due interact residu choos amino acid type posit constrain possibl order result protein proper structur activ account depend residu advantag independ develop approach protein design base undirect probabilist graphic model fig approach construct graphic model encod residu constraint model gener produc sequenc optim meet constraint focu constraint due residu coupl common pair amino acid type pair posit correl mutat evolv residu recent ranganathan colleagu show account residu coupl addit conserv extent suffici viabil ww domain previous develop approach learn undirect graphic model encapsul conserv coupl constraint protein famili model formal probabilist semant reason amino acid choic defin probabl distribut function measur sequenc satisfi coupl constraint observ extant sequenc famili order design contact replac nmr reson assign motiv complement tradit role structur studi protein nuclear magnet reson nmr spectroscopi plai increasingli import role function studi nmr dynam experi character motion involv target recognit ligand bind nmr chemic shift perturb experi identifi local protein protein protein ligand interact kei bottleneck studi determin backbon reson assign spectral peak map specif atom paper develop approach address bottleneck exploit rai structur homolog model assign entir backbon set fast cheap nmr experi result formul contact replac reson assign problem comput correspond contact graph repres structur nmr graph repres data nmr graph significantli corrupt ambigu version contact graph show combin connect amino acid type inform exploit random structur nois provabl determin uniqu correspond polynomi time high probabl presenc signific nois constant number noisi edg vertex detail effici random algorithm show varieti experiment synthet dataset robust typic level structur variat nois miss algorithm achiev good assign accuraci helic sheet loop region avail contact replac algorithm implement platform independ python code softwar freeli obtain academ request author contact algorithm joint optim stabil divers plan combinatori librari chimer protein abstract engin protein variant construct screen combinatori librari chimer protein complementari compet goal desir protein similar evolutionarili select wild type protein stabli fold displai function variat present method stavers simultan optim stabil divers select set breakpoint locat site direct recombin goal uncov undomin breakpoint set breakpoint set factor algorithm find undomin set serv vertic lower envelop dimension stabil divers convex hull breakpoint set algorithm identifi addit breakpoint set concav undomin domin undiscov breakpoint set distanc bound comput algorithm algorithm effici requir gaussian process model spatial data mine graphic model residu coupl protein famili identifi residu coupl relationship protein famili provid import insight famili evolutionari record signific applic analyz optim sequenc structur function relationship present algorithm infer undirect graphic model repres residu coupl protein famili model call residu coupl network serv compact descript joint amino acid distribut focus independ residu stand contrast current method manipul dens represent variat focus assess depend conflat direct indirect relationship probabilist model sound basi predict newli design protein fold function diagnost protein stabl function abduct reason attempt graft featur protein famili algorithm readili incorpor prior hypothes underli mechanist energet explan coupl result approach constitut power discriminatori mechan identifi residu coupl protein sequenc structur analysi result protein coupl receptor gpcr pdz domain famili demonstr abil approach effect uncov exploit model residu coupl structur discoveri massiv spatial data set intellig simul tool hypergraph model multi dresidu interact protein sequenti dconstrain partit algorithm optim site ddirect protein recombin abstract relationship amino acid determin stabil function constrain evolutionari histori develop probabilist hypergraph model residu relationship gener tradit pairwis contact potenti account statist multi residu interact model detect random associ protein famili protein databas model optim site direct recombin experi preserv signific interact increas frequenc gener recombin formul optim sequenti constrain hypergraph partit problem qualiti recombin librari wrt set breakpoint character total perturb edg weight prove problem np hard gener develop exact heurist polynomi time algorithm number import case applic beta lactamas famili demonstr util algorithm plan site direct recombin hypergraph model multi dresidu interact protein sequenti dconstrain partit algorithm optim site ddirect protein recombin abstract relationship amino acid determin stabil function constrain evolutionari histori develop probabilist hypergraph model residu relationship gener tradit pairwis contact potenti account statist multi residu interact model detect random associ protein famili protein databas model optim site direct recombin experi preserv signific interact increas frequenc gener recombin formul optim sequenti constrain hypergraph partit problem qualiti recombin librari wrt set breakpoint character total perturb edg weight prove problem np hard gener develop exact heurist polynomi time algorithm number import case applic beta lactamas famili demonstr util algorithm plan site direct recombin effici random algorithm contact dbase nmr backbon reson assign motiv backbon reson assign critic bottleneck studi protein structur dynam interact nuclear magnet reson nmr spectroscopi minimalist approach assign call contact base seek dramat reduc experiment time expens replac standard suit bond experi space noesi experi contact base approach spectral data repres graph vertic put residu unknown relat primari sequenc edg hypothes noesi interact observ spectral peak explain residu close due experiment ambigu incorrect edg hypothes spectral peak assign deriv identifi consist pattern edg helic sheet graph map spatial data mine support pandem prepared reconsid complet search algorithm protein backbon nmr assign gaussian process activ data mine spatial aggreg present activ data mine mechan qualit analysi spatial dataset integr identif analysi structur spatial data target collect addit sampl mechan design spatial aggreg languag sal qualit spatial reason seek uncov high level spatial structur spars set sampl approach import applic domain aircraft design wireless system simul fluid dynam sensor network mechan emploi gaussian process formal mathemat model reason spatial data order build surrog model spars data reason uncertainti estim unsampl point formul object criteria close loop data collect data analysi optim sampl select entropi base function defin spatial aggreg tradit approach sampl minim estim varianc appli mechan global optim benchmark compris testbank function data wireless system simul result reveal propos sampl strategi make judici data point select locat clarifi high level structur data choos point improv qualiti function approxim random graph approach nmr sequenti assign nuclear magnet reson nmr spectroscopi scientist studi protein structur dynam interact solut step applic determin reson assign map spectral data atom residu primari sequenc autom reson assign algorithm reli inform connect bond atom interact amino acid type typic determin string connect residu map string posit primari sequenc signific ambigu exist connect amino acid type inform paper focus inform content connect develop random graph theoret framework algorithm connect driven nmr sequenti assign random graph model captur structur chemic shift degeneraci kei sourc connect ambigu give simpl natur random algorithm find optim assign set connect fragment nmr graph algorithm natur effici reus substr explor connect choic overcom local ambigu enforc global consist choic analyz algorithm random graph model show provabl toler larg ambigu give expect optim perform polynomi time present result practic applic algorithm experiment dataset varieti protein experiment set up demonstr approach overcom signific nois local ambigu identifi signific fragment sequenti assign kei word nuclear magnet reson nmr spectroscopi autom sequenti reson assign random graph model random algorithm hamiltonian path spatial aggreg qualit assess scientif comput qualit assess scientif comput emerg applic area appli data driven approach character high level phenomena includ condit matric sensit type error propag algorithm converg behavior paper develop spatial aggreg approach formal analysi term model select util spatial structur extract matrix perturb dataset focu character matrix eigenstructur analyz sensit comput spectral portrait determin eigenvalu multipl jordan portrait approach emploi spatial reason overcom nois sparsiti detect mutual reinforc interpret guid subsequ data sampl enabl quantit evalu properti scientif comput term confid model explain term sampl data domain knowledg underli mathemat structur methodolog rigor common approach visual inspect substanti effici due defin stop criteria result show mechan effici sampl perturb space successfulli uncov high level properti matric geometr analysi cross dlinkabl protein fold discrimin protein structur insight evolutionari origin function mechan protein pursu minimalist approach protein fold identif character fold term consist geometr featur restraint deriv cheap highthroughput experi experi residu specif cross link analyz mass spectrometri paper present suit lower upper bound algorithm analyz distanc surfac cross link site valid predict model experiment cross link result analysi comput experi simul publish experiment data demonstr algorithm enabl effect model discrimin random graph approach nmr sequenti assign nuclear magnet reson nmr spectroscopi scientist studi protein structur dynam interact solut step applic determin reson assign map spectral data atom residu primari sequenc autom reson assign algorithm reli inform connect bond atom interact amino acid type typic determin string connect residu map string posit primari sequenc signific ambigu exist connect amino acid type inform paper focus inform content connect develop random graph theoret framework algorithm connect driven nmr sequenti assign random graph model captur structur chemic shift degeneraci kei sourc connect ambigu give simpl natur random algorithm find optim assign set connect fragment nmr graph algorithm natur effici reus substr explor connect choic overcom local ambigu enforc global consist choic analyz algorithm random graph model show provabl toler larg ambigu give expect optim perform polynomi time present result practic applic algorithm experiment dataset varieti protein experiment set up demonstr approach overcom signific nois local ambigu identifi signific fragment sequenti assign kei word nuclear magnet reson nmr spectroscopi autom sequenti reson assign random graph model random algorithm hamiltonian path qualit spatial reason extract reason spatial aggreg reason spatial data kei task applic includ geograph inform system meteorolog fluid flow analysi comput aid design protein structur databas applic requir identif manipul qualit spatial represent detect object occlud digit imag effici determin relationship propos road wetland region geograph data set qualit spatial reason qsr represent primit spatial vocabulari infer mechan task paper review repres work qsr data poor scenario goal design represent answer qualit queri numer inform turn data rich case goal deriv manipul qualit spatial represent effici correctli abstract import spatial aspect underli data subsequ task paper focus qsr system spatial aggreg sa answer spatial queri scientif engin data set case studi applic weather analysi illustr effect represent reason support data poor data rich form qsr servic learn softwar engin mainten gaussian process model spatial aggreg algorithm multi level spatial aggreg import data mine varieti scientif engin applic analysi weather data aggreg temperatur pressur data ridg front perform analysi wireless system aggreg simul result configur space region exhibit perform characterist applic data collect expens time consum effort focus gather sampl locat import analysi requir function model data mine algorithm order assess impact potenti sampl mine suitabl spatial aggreg paper describ gaussian process approach model multi layer spatial aggreg algorithm demonstr abil result model captur essenti underli qualit behavior algorithm help cast classic spatial aggreg algorithm rigor quantit framework gaussian process model support divers direct sampl character sensit mine algorithm paramet understand variat input data field percol spatial aggreg hierarchi physic dbase encapsul embed softwar distribut sens control applic spatial aggreg abstract data aris distribut embed sens control applic set call spatio tempor object local continu underli physic problem domain give rise spatial coher tempor contigu object metric space parameter physic properti locat intens light temperatur pressur motion veloc object aggreg abstract abstract descript applic written creation transform abstract object qualit analysi correspond experiment algorithm sampl strategi mine data dscarc domain descript redescrib redescrib updat object graph class equival spatial figur spatial aggreg languag multilay spatial aggreg uncov uniform vocabulari oper domain knowledg express scientif data mine task vector field bundl contour aggreg correspond abstract cluster uncov region uniform multilevel comput sal aggreg hierarchi streamlin bundl converg flow salplcpmk compmkm streamlin aggreg neighborhood grap choos basi curvatur match direct vector aggreg data scarc classif decis ambigu multipp streamlin aggreg exist case choos datasampm reduc ambigu clarifi correct classif essenc oursampl methodolog sal aggreg identifi inform theoret measur ambigu drive stage futur data collect instanc summar ambigu streamlin classif ambigu distribut aspal locat detect ambigu ambigu reduct apmeel minim maxim case function involv thecompnkn ambigu function underli data field sentrop ambigu distribut reveal minim lead select data ptac clarifi distribut streamlin effect data data miningpningclr methodolog net effect cancaplbm appmbpcdpi design desi physic geometr algorithm structur molecular biologi wealth interest comput problem aris propos method discov pharmaceut paper survei recent work kei area physic geometr algorithm pga approach data interpret experi plan drug design data direct comput protocol high throughput protein structur determin kei compon structur determin nuclear magnet reson nmr assign spectral peak develop approach call jigsaw autom secondari structur determin main chain assign jigsaw consist main compon graph base secondari structur pattern identif unassign heteronuclear label nmr data assign spectral peak probabilist align identifi secondari structur element primari sequenc experi plan data interpret algorithm reduc mass degeneraci mass spectrometri ms ms offer advantag highthroughput assai small sampl size larg mass limit face potenti problem mass degeneraci indistinguish mass multipl biopolym fragment limit proteolyt digest studi select isotop label substanti reduc potenti mass degeneraci context structur determin protein protein protein dna complex ambigu ddirect sampl qualit analysi spars data spatial ddistribut physic system number import scientif engin influenc dbase model decomposit reason spatial distribut physic system distribut semiconductor wafer control nois photocopi machin requir interpret distribut data design decentr control spatial distribut system develop effect comput techniqu repres reason system model partial differenti equat pde major challeng problem qualit spatial reason research paper introduc approach decentr control design influencebas model decomposit appli context thermal regul influenc base model decomposit decentr model call influenc graph kei data abstract repres influenc control distribut physic field serv basi algorithm control placement paramet design distribut system larg number coupl variabl algorithm exploit physic knowledg local linear superpos continu encapsul influenc graph repres depend field node control node control placement design algorithm util influenc graph decompos problem domain decoupl result region decentr control paramet optim algorithm util influenc graph effici evalu thermal field explicitli trade comput commun control qualiti leverag physic knowledg encapsul influenc graph control design algorithm effici standard techniqu produc design explain term problem structur reduc mass degeneraci sar ms stabl isotop label mass spectrometri ms promis invalu tool function genom support low cost high throughput experi larg scale ms face potenti problem mass degeneraci indistinguish mass multipl biopolym fragment limit proteolyt digest paper studi task plan interpret ms experi select isotop label substanti reduc potenti mass degeneraci algorithm support experiment comput protocol call structur activ relat mass spectrometri sar ms elucid function protein dna protein protein complex sar ms enzymat cleav crosslink complex analyz result mass spectrum mass peak hypothes fragment depend bind mode cleavag site shield absenc anticip peak implic fragment part interact region inaccess due conform chang bind mass spectra provid evid structur activ relat address combinatori algorithm question area data analysi constrain bind mode base mass signatur experi plan determin isotop label strategi reduc mass degeneraci aid data analysi explor comput complex problem obtain upper lower bound report experiment result implement algorithm reduc mass degeneraci sar ms stabl isotop label mass spectrometri ms promis invalu tool function genom support low cost high throughput experi larg scale ms face potenti problem mass degeneraci indistinguish mass multipl biopolym fragment limit proteolyt digest paper studi task plan interpret ms experi select isotop label substanti reduc potenti mass degeneraci algorithm support experiment comput protocol call structur activ relat mass spectrometri sar ms elucid function protein dna protein protein complex sar ms enzymat cleav crosslink complex analyz result mass spectrum mass peak hypothes fragment depend bind mode cleavag site shield absenc anticip peak implic fragment part interact region inaccess due conform chang bind mass spectra provid evid structur activ relat address combinatori algorithm question area data analysi constrain bind mode base mass signatur experi plan determin isotop label strategi reduc mass degeneraci aid data analysi explor comput complex problem obtain upper lower bound report experiment result implement algorithm influenc dbase model decomposit recent rapid advanc mem inform process technolog enabl gener ai robot system call smart matter system sensor rich physic embed system rang decentr control system regul build temperatur smart build vehicl board diagnost control system interrog larg amount sensor data core task construct oper smart matter system synthes optim control polici data rich model system environ model thousand coupl real valu variabl prohibit expens reason tradit optim techniqu neural net genet algorithm paper introduc gener mechan automat decompos larg model smaller subpart subpart separ optim combin intellig simul tool mine larg scientif data set paper describ problem challeng opportun intellig simul physic system prototyp intellig simul tool construct interpret massiv data set physic field design engin system identifi characterist intellig simul describ concret applic exampl applic includ weather data interpret distribut control optim spatio tempor diffus reaction pattern analysi demonstr intellig simul tool indispens rapid prototyp applic program challeng scientif engin domain qualit analysi distribut physic system applic control synthesi import physic phenomena temperatur distribut air flow acoust wave continu distribut paramet field analyz control physic process system common task scientif engin domain challeng multifold distribut field conceptu harder reason lump paramet model comput method prohibit expens complex spatial domain underli physic impos sever constraint observ control paper develop ontolog abstract structur base design mechan framework collect spatial aggreg sa reason synthes distribut control scheme physic field ontolog abstract model physic field hierarchi network spatial object sa appli small number gener oper field comput concis structur spatial aggreg languag applic spatial aggreg framework organ comput imag analogu represent physic process data interpret control task conceptu common comput structur class implement problem solver difficult scientif engin problem compris mechan languag program style spatial aggreg mechan transform numer input field success higher level descript appli small ident set oper layer metric neighborhood relat equival relat paper describ spatial aggreg languag applic spatial aggreg languag abstract data type neighborhood graph field set interfac oper construct transform field librari compon implement user mix match special ap 