Lisebelisoa tse kholo li ntlafalitse k'hemistri e kholo ka 2022
Disete tse kgolo tsa data le disebediswa tse kgolo di thusitse bo-rasaense ho sebetsana le k'hemistri ka tekanyo e kgolo selemong sena
kaAriana Remmel
Mokitlane: Setsi sa Khomphutha sa Boetapele ba Oak Ridge ho ORNL
Khomphutha e kgolo ya Frontier Laboratoring ya Naha ya Oak Ridge ke ya pele molokong o motjha wa mechine e tla thusa bo-ramakhemistri ho nka di-simulation tsa dimolek'hule tse rarahaneng ho feta leha e le neng pele.
Bo-rasaense ba entse litšibollo tse kholo ka lisebelisoa tse kholo ka 2022. Ha ba haha holim'a mokhoa oa morao tjena oa bohlale ba maiketsetso bo nang le bokhoni ba lik'hemik'hale, bafuputsi ba entse tsoelo-pele e kholo, ba ruta lik'homphieutha ho bolela esale pele meaho ea protheine ka tekanyo e sa tloaelehang. Ka Phupu, k'hamphani ea DeepMind e nang le Alphabet e phatlalalitse database e nang le meaho eahoo e ka bang liprotheine tsohle tse tsejoang—Liprotheine tse fetang limilione tse 200 tse tsoang mefuteng e fetang limilione tse 100—joalokaha ho boletsoe esale pele ke algoriti ea ho ithuta ka mochini AlphaFold. Eaba, ka Pulungoana, k'hamphani ea theknoloji Meta e bontšitse tsoelo-pele ea eona theknolojing ea ho bolela esale pele protheine ka algoriti ea AI e bitsoangESMFold. Phuputsong ea pele ho khatiso e e-so hlahlojoe ke lithaka, bafuputsi ba Meta ba tlalehile hore mokhoa oa bona o mocha oa ho etsa lintho ha o nepahale joalo ka AlphaFold empa o potlakile. Lebelo le eketsehileng le ne le bolela hore bafuputsi ba ka bolela esale pele meaho e limilione tse 600 ka libeke tse 2 feela (bioRxiv 2022, DOI:10.1101/2022.07.20.500902).
Litsebi tsa baeloji Sekolong sa Bongaka sa Univesithi ea Washington (UW) lia thusaatolosa bokgoni ba dikhomphutha ba dikhemikhale ho feta template ya tlhahoka ho ruta mechine ho hlahisa liprotheine tse etselitsoeng ho tloha qalong. David Baker oa UW le sehlopha sa hae ba thehile sesebelisoa se secha sa AI se ka ralang liprotheine ka ho ntlafatsa ka makhetlo-khetlo ka litlhahiso tse bonolo kapa ka ho tlatsa likheo lipakeng tsa likarolo tse khethiloeng tsa sebopeho se seng se ntse se le teng (Saense2022, DOI:10.1126/saense.abn2100Sehlopha se boetse se qalile lenaneo le lecha, ProteinMPNN, le ka qalang ho tloha libopehong le likopanong tse entsoeng ka 3D tsa li-subunit tse ngata tsa protheine ebe le fumana tatellano ea li-amino acid tse hlokahalang ho li etsa ka katleho (Saense2022, DOI:10.1126/saense.add2187;10.1126/saense.add1964). Li-algorithm tsena tse nang le tsebo ea lik'hemik'hale li ka thusa bo-rasaense ho aha meralo ea liprotheine tsa maiketsetso tse ka sebelisoang linthong tse ncha tsa biomaterial le meriana.
Mokitlane: Ian C. Haydon/UW Institute for Protheine Design
Mekhoa ea ho ithuta ka mochini e thusa bo-rasaense ho lora liprotheine tse ncha ba nahanne ka mesebetsi e itseng.
Ha litabatabelo tsa litsebi tsa lik'hemistri tsa k'homphieutha li ntse li hola, lik'homphieutha tse sebelisoang ho etsisa lefats'e la limolek'hule le tsona lia hola. Ho Oak Ridge National Laboratory (ORNL), litsebi tsa lik'hemistri li ile tsa bona e 'ngoe ea li-supercomputer tse matla ka ho fetisisa tse kileng tsa hahoa.Khomphutha e kholo ea ORNL, Frontier, ke e 'ngoe ea mechini ea pele e balang mesebetsi e fetang quintillion e le 'ngoe e phaphametseng ka motsotsoana, yuniti ea lipalo tsa k'homphieutha. Lebelo leo la k'homphieutha le potlakile ka makhetlo a mararo ho feta 'mampoli ea busang, Fugaku ea supercomputer Japane. Selemong se tlang, lilaboratori tse ling tse peli tsa naha li rera ho qala lik'homphieutha tsa exascale US. Matla a k'homphieutha a fetang a mechini ena ea sejoale-joale a tla lumella litsebi tsa k'hemistri ho etsisa litsamaiso tse kholo le ho feta linako tse telele. Lintlha tse bokelletsoeng ho tsoa mehlaleng eo li ka thusa bafuputsi ho sutumelletsa meeli ea se ka khonehang k'hemistring ka ho fokotsa lekhalo pakeng tsa likarabo ka har'a flask le li-simulation tsa sebele tse sebelisoang ho li etsa mohlala. "Re fihlile ntlheng eo re ka qalang ho botsa lipotso mabapi le hore na ke eng e haellang mekhoeng ea rona ea khopolo-taba kapa mehlala e tla re atametsa haufi le seo teko e re bolellang sona e le 'nete," Theresa Windus, setsebi sa k'hemistri sa k'hemistri Univesithing ea Naha ea Iowa le moetapele oa projeke le Exascale Computing Project, o boleletse C&EN ka Loetse. Diketsiso tse sebediswang dikhomphutheng tsa exascale di ka thusa bo-ramakhemistri ho qapa mehlodi e metjha ya mafura le ho rala thepa e ntjha e mamellang boemo ba lehodimo.
Ho pholletsa le naha, Menlo Park, California, SLAC National Accelerator Laboratory e ntse e kenngwaLintlafatso tse ntle haholo ho ea ho Mohloli oa Leseli o Kopaneng oa Linac (LCLS)seo se ka dumella bo-ramakhemistri ho sheba ka botebo lefatsheng la diathomo le dielektrone tse potlakileng ka ho fetisisa. Setsi sena se hahiloe hodima accelerator e otlolohileng ya 3 km, eo dikarolo tsa yona di phodisitsweng ka helium e metsi ho fihlela ho 2 K, ho hlahisa mofuta wa mohlodi wa lesedi o matla ka ho fetisisa o bitswang X-ray free-electron laser (XFEL). Bo-ramakhemistri ba sebedisitse di-pulse tse matla tsa disebediswa ho etsa difilimi tsa dimolek'hule tse ba thusitseng ho shebella ditshebetso tse ngata, tse kang ho bopa dibonto tsa dikhemikhale le dienzyme tsa photosynthetic tse tla sebetsa. “Ka motsotsoana wa femtosecond, o ka bona diathomo di eme, dibonto tsa athomo tse le nngwe di robeha,” Leora Dresselhaus-Marais, rasaense wa thepa ya nang le kopano e kopanetsweng Univesithing ya Stanford le SLAC, o boleletse C&EN ka Phupu. Dintlafatso tsa LCLS di tla boela di dumella bo-ramahlale ho ntlafatsa matla a di-X-ray ha bokgoni bo botjha bo fumaneha mathoasong a selemo se tlang.
Mokitlane: Laboratori ea Naha ea Accelerator ea SLAC
Lese ea X-ray ea SLAC National Accelerator Laboratory e hahiloe holim'a accelerator e otlolohileng ea 3 km Menlo Park, California.
Selemong sena, bo-rasaense ba boetse ba bone hore na Telescope ea Sepakapaka ea James Webb (JWST) e neng e letetsoe ka nako e telele e ka ba matla hakae bakeng sa ho senolaho rarahana ha lik'hemik'hale tsa bokahohle ba ronaNASA le balekane ba yona—European Space Agency, Canadian Space Agency, le Space Telescope Science Institute—ba se ba lokolotse ditshwantsho tse ngata, ho tloha ditshwantshong tse kganyang tsa dinaledi tse nang le mahlasedi a manyane ho isa ho menwana ya dinaledi tsa boholo-holo. Telescope ya infrared ya $10 bilione e kgabisitswe ka disuite tsa disebediswa tsa saense tse reretsweng ho hlahloba nalane e tebileng ya bokahohle ba rona. Ka dilemo tse mashome e ntse e etswa, JWST e se e fetile ditebello tsa baenjiniere ba yona ka ho nka setshwantsho sa sehlopha sa dinaledi se potolohang kamoo se neng se hlaha kateng dilemong tse dibilione tse 4.6 tse fetileng, se feletse ka ditshwaelo tsa spectroscopic tsa oksijene, neon le diathomo tse ding. Bo-rasaense ba boetse ba lekantse ditshwaelo tsa maru a nang le mouoane le leru polaneteng ya exoplanet, ba fana ka data e ka thusang ditsebi tsa dinaledi ho batla mafatshe a ka dulwang ka nqane ho Lefatshe.
Nako ea poso: Hlakola-07-2023