Kuthiwani uma singasebenzisa ubuhlakani bokwenziwa ukuze siphendule enye yezimfihlakalo ezinkulu zokuphila - ukugoqa amaprotheni? Sekungamashumi eminyaka ososayensi besebenza kulokhu.
Imishini manje ingabikezela ukwakheka kwamaprotheni ngokunemba okumangalisayo isebenzisa amamodeli okufunda ajulile, ukuguqula ukuthuthukiswa kwezidakamizwa, i-biotechnology, kanye nolwazi lwethu lwezinqubo ezibalulekile zebhayoloji.
Ngijoyine ekuhloleni indawo ekhangayo yokugoqa amaprotheni e-AI, lapho ubuchwepheshe obusezingeni eliphezulu bushayisana nobunkimbinkimbi bempilo ngokwayo.
Ukwembula Imfihlakalo Yokugoqa Amaprotheni
Amaprotheni asebenza emizimbeni yethu njengemishini emincane ukwenza imisebenzi ebalulekile njengokuphula ukudla noma ukuthutha umoya-mpilo. Kumele zigoqwe ngendlela efanele ukuze zisebenze kahle, njengoba nje nokhiye kufanele usikwe ngendlela efanele ukuze ungene kungidi. Ngokushesha nje lapho amaprotheni edalwa, inqubo yokugoqa eyinkimbinkimbi kakhulu iqala.
Ukugoqa amaprotheni kuyinqubo lapho amaketanga amade ama-amino acid, amabhlogo wokwakha amaprotheni, agoqeka abe izingxenye ezintathu-ntathu ezilawula ukusebenza kwephrotheni.
Cabangela intambo ende yobuhlalu okufanele i-odwe ngendlela enembile; yilokhu okwenzeka lapho iphrotheni igoqa. Nokho, ngokungafani nobuhlalu, ama-amino acid anezici eziyingqayizivele futhi asebenzelana ngezindlela ezihlukahlukene, okwenza ukugoqa amaprotheni kube inqubo eyinkimbinkimbi nebucayi.
Isithombe esilapha simelela i-hemoglobin yomuntu, okuyiprotheni egoqiwe eyaziwa kakhulu
Amaprotheni kufanele agoqe ngokushesha nakancane, kungenjalo azosongeka futhi abe nesici. Lokho kungaholela ezifweni ezifana ne-Alzheimer kanye ne-Parkinson. Izinga lokushisa, ukucindezela, nokuba khona kwamanye ama-molecule engqamuzaneni konke kunomthelela enqubweni yokugoqa.
Ngemva kwamashumi eminyaka okucwaninga, ososayensi basazama ukuthola ukuthi amaprotheni agoqeka kanjani.
Ngokujabulisayo, intuthuko kwezobuhlakani bokwenziwa ithuthukisa intuthuko kulo mkhakha. Ososayensi bangabikezela ukwakheka kwamaprotheni ngokunembe kunanini ngaphambili ngokusebenzisa umshini wokufunda ama-algorithms ukuhlola umthamo omkhulu wedatha.
Lokhu kunamandla okushintsha ukuthuthukiswa kwemithi futhi kwandise ulwazi lwethu lwamangqamuzana ngalesi sifo.
Ingabe Imishini Isebenza Kangcono?
Izindlela Ezivamile Zokusonga Amaprotheni Zinemikhawulo
Sekungamashumi eminyaka ososayensi bezama ukuthola ukugoqeka kwamaprotheni, kodwa ubunkimbinkimbi benqubo buye benza lesi sihloko saba inselele.
Izindlela ezivamile zokubikezela ukwakheka kwamaprotheni zisebenzisa inhlanganisela yezindlela zokuhlola kanye nokumodela kwekhompyutha, nokho, lezi zindlela zonke zinezihibe.
Amasu okuhlola afana ne-X-ray crystallography kanye ne-nuclear magnetic resonance (NMR) angadla isikhathi futhi abize. Futhi, amamodeli ekhompiyutha ngezinye izikhathi ancike ekucabangeni okulula, okungaholela ekubikezelweni okuyiphutha.
I-AI Ingazinqoba Lezi Zithiyo
Ngenhlanhla, ukuhlakanipha okungekhona okwangempela inikeza isithembiso esisha sokubikezela okunembe kakhudlwana kwesakhiwo samaprotheni. Ama-algorithm okufunda komshini angahlola umthamo omkhulu wedatha. Futhi, bathola amaphethini abantu abangawageja.
Lokhu kube nomphumela wokwakhiwa kwamathuluzi esofthiwe amasha nezinkundla ezikwazi ukubikezela ukwakheka kwamaprotheni ngokunemba okungenakuqhathaniswa.
Ama-algorithms Wokufunda Womshini Othembisa Kakhulu We-Protein Structure Prediction
Isistimu ye-AlphaFold eyakhiwe ngabakwa-Google Deepmind iqembu lingenye yentuthuko ethembisa kakhulu kule ndawo. Ithole intuthuko enkulu eminyakeni yamuva nje ngokusebenzisa ubuchule bokufunda obujulile ukubikezela ukwakheka kwamaprotheni ngokusekelwe ngokulandelana kwawo kwe-amino acid.
Amanethiwekhi e-Neural, imishini esekela i-vector, namahlathi angahleliwe aphakathi kwezindlela zokufunda zomshini ezibonisa isithembiso sokubikezela ukwakheka kwamaprotheni.
Lawa ma-algorithms angafunda kumadathasethi amakhulu. Futhi, bangalindela ukuhlobana phakathi kwama-amino acid ahlukahlukene. Ngakho, ake sibone ukuthi kusebenza kanjani.
Ukuhlaziya okuhlangene kokuziphendukela kwemvelo kanye nesizukulwane sokuqala se-AlphaFold
Impumelelo I-AlphaFold yakhelwe phezu kwemodeli yenethiwekhi ye-neural ejulile eyasungulwa kusetshenziswa ukuhlaziywa kokuziphendukela kwemvelo. Umqondo wokuziphendukela kwemvelo uthi uma ama-amino acid amabili ephrotheni ehlangana namanye, azokhula ndawonye ukuze agcine ukuxhumana kwawo okusebenzayo.
Abacwaningi bangathola ukuthi yimaphi amapheya ama-amino acid okungenzeka ukuthi athinteke esakhiweni se-3D ngokuqhathanisa ukulandelana kwe-amino acid yamaphrotheni amaningi afanayo.
Le datha isebenza njengesisekelo sokuphindaphinda kokuqala kwe-AlphaFold. Ibikezela ubude phakathi kwamapheya e-amino acid kanye nama-engeli amabhondi e-peptide awaxhumayo. Le ndlela yayizidlula zonke izindlela zangaphambili zokubikezela ukwakheka kwamaprotheni kusukela ekulandelaneni, nakuba ukunemba kwakusakhawulelwe kumaphrotheni angenazo izifanekiso ezibonakalayo.
I-AlphaFold 2: Indlela Entsha Kakhulu
I-AlphaFold2 isofthiwe yekhompuyutha eyakhiwe i-DeepMind esebenzisa ukulandelana kwe-amino acid yephrotheni ukubikezela ukwakheka kwe-3D kwephrotheni.
Lokhu kubalulekile ngoba ukwakheka kwephrotheni kuyayisho indlela esebenza ngayo, futhi ukuqonda umsebenzi wayo kungasiza ososayensi bakhe imithi eqondise iphrotheni.
Inethiwekhi ye-neural ye-AlphaFold2 ithola njengokufakiwe ukulandelana kwe-amino acid yeprotheni kanye nemininingwane yokuthi lokho chungechunge kuqhathaniswa kanjani nokunye ukulandelana kusizindalwazi (lokhu kubizwa ngokuthi “ukuqondanisa ukulandelana”).
Inethiwekhi ye-neural yenza isibikezelo mayelana nesakhiwo se-3D sephrotheni ngokusekelwe kulokhu okokufaka.
Yini Ehlukanisa Ku-AlphaFold2?
Ngokungafani nezinye izindlela, i-AlphaFold2 ibikezela ukwakheka kwe-3D kwangempela kwephrotheni kunokumane nje kuhlukaniswe phakathi kwamapheya ama-amino acid noma ama-engeli phakathi kwamabhondi awaxhumayo (njengoba kwenziwa ama-algorithms angaphambilini).
Ukuze inethiwekhi ye-neural ilindele ukwakheka okuphelele ngesikhathi esisodwa, isakhiwo sibhalwa ngekhodi ekupheleni ukuya ekupheleni.
Esinye isici esibalulekile se-AlphaFold2 ukuthi inikeza isilinganiso sokuthi iqiniseka kangakanani kwisibikezelo sayo. Lokhu kuvezwa njengombhalo wekhodi esakhiweni esilindelwe, okubomvu okumelela ukuzethemba okuphezulu nokuluhlaza okwesibhakabhaka okuphakamisa ukuzethemba okuphansi.
Lokhu kuyasiza ngoba kwazisa ososayensi ngokuzinza kwesibikezelo.
Ukubikezela Isakhiwo Esihlanganisiwe Sokulandelana Okuningana
Ukunwetshwa kwakamuva kwe-Alphafold2, okwaziwa nge-Alphafold Multimer, kubikezela ukwakheka okuhlanganisiwe kokulandelana okuningi. Isenamaphutha aphezulu ngisho noma isebenza kangcono kakhulu kunamasu angaphambili. Amaphesenti angama-25 nje angama-4500 wezakhiwo zamaprotheni abikezelwe ngempumelelo.
U-70% wezindawo ezimahhadlahhadla zokwenza ukuthintana wabikezelwa ngendlela efanele, kodwa ukuma okuhlobene kwamaphrotheni amabili kwakungalungile. Uma ukujula kokuqondanisa okumaphakathi kungaphansi kokulandelana okungaba ngu-30, ukunemba kokuqagela kwe-Alphafold multimer kwehla kakhulu.
Ungazisebenzisa kanjani izibikezelo ze-Alphafold
Amamodeli abikezelwe asuka ku-AlphaFold anikezwa ngamafomethi wefayela afanayo futhi angasetshenziswa ngezindlela ezifanayo nezakhiwo zokuhlola. Kubalulekile ukucabangela izilinganiso zokunemba ezinikezwa nemodeli ukuze kugwenywe ukungezwani.
Kuyasiza ikakhulukazi ezakhiweni eziyinkimbinkimbi njengama-homomer ahlanganisiwe noma amaprotheni agoqa kuphela lapho kukhona
i-ligand engaziwa.
Ezinye Izinselele
Inkinga enkulu ekusebenziseni izakhiwo ezibikezelwe ukuqonda ukuguquguquka, ukukhetha kwe-ligand, ukulawula, i-allostery, izinguquko zangemuva kokuhumusha, kanye ne-kinetics yokubopha ngaphandle kokufinyelela kumaprotheni kanye nedatha ye-biophysical.
Ukufunda komshini kanye nocwaningo lwe-physics-based molecular dynamics lungasetshenziswa ukunqoba le nkinga.
Lolu phenyo lungase luhlomule ekwakhiweni kwekhompyutha okuyisipesheli nangempumelelo. Nakuba i-AlphaFold izuze intuthuko enkulu ekubikezeleni izakhiwo zamaprotheni, kusekuningi okufanele kufundwe emkhakheni we-structural biology, futhi izibikezelo ze-AlphaFold ziyisiqalo socwaningo oluzayo.
Yimaphi Amanye Amathuluzi Aphawulekayo?
I-RoseTTFold
I-RoseTTAFold, eyakhiwe abacwaningi baseNyuvesi yaseWashington, ngokufanayo isebenzisa ama-algorithms okufunda ajulile ukuze ibikezele izakhiwo zamaprotheni, kodwa futhi ihlanganisa indlela yenoveli eyaziwa ngokuthi “i-torsion angle dynamics simulations” ukuze ithuthukise izakhiwo ezibikezelwe.
Le ndlela ibe nemiphumela ekhuthazayo futhi ingase ibe usizo ekunqobeni imikhawulo yamathuluzi okugoqa amaprotheni e-AI akhona.
iRosetta
Elinye ithuluzi, i-trRosetta, libikezela ukugoqa kwamaprotheni ngokusebenzisa i-a inethiwekhi ye-neural uqeqeshwe ngezigidi zokulandelana kwamaprotheni kanye nezakhiwo.
Iphinde isebenzise indlela “yokumodela okusekelwe kusifanekiso” ukuze idale ukubikezela okunembe kakhudlwana ngokuqhathanisa iphrotheni eqondiwe nezakhiwo ezaziwayo ezifanayo.
Kubonisiwe ukuthi i-trRosetta iyakwazi ukubikezela ukwakheka kwamaphrotheni amancane namaphrotheni.
I-DeepMetaPSICOV
I-DeepMetaPSICOV ingelinye ithuluzi eligxile ekubikezeleni amamephu okuxhumana namaprotheni. Lezi, zisetshenziswa njengomhlahlandlela wokubikezela ukugoqa kwamaprotheni. Isebenzisa ukufunda okujulile izindlela zokubikezela amathuba okusebenzelana kwezinsalela ngaphakathi kwephrotheni.
Lezi zisetshenziselwa ukubikezela imephu yokuxhumana iyonke. I-DeepMetaPSICOV ibonise amandla ekubikezeleni ukwakheka kwamaprotheni ngokunemba okukhulu, noma ngabe izindlela zangaphambilini zehlulekile.
Likuphatheleni Ikusasa?
Ikusasa lokugoqa amaprotheni e-AI liqhakazile. Ama-algorithms asekelwe ekufundeni okujulile, ikakhulukazi i-AlphaFold2, muva nje enze inqubekelaphambili enkulu ekubikezeleni ngokuthembekile izakhiwo zamaprotheni.
Lokhu okutholakele kunamandla okuguqula ukuthuthukiswa kwezidakamizwa ngokuvumela ososayensi ukuthi baqonde kangcono ukwakheka nokusebenza kwamaprotheni, okuyizinhloso ezivamile zokwelapha.
Noma kunjalo, izindaba ezinjengokubikezela izakhi zamaprotheni kanye nokuthola isimo sangempela sokusebenza sezakhiwo okulindelekile zisekhona. Kudingeka ucwaningo olwengeziwe ukuze kuxazululwe lezi zinkinga futhi kwandiswe ukunemba nokuthembeka kwe-AI protein algorithms.
Nokho, izinzuzo ezingaba khona zalobu buchwepheshe zinkulu, futhi bunamandla okuholela ekukhiqizweni kwemithi esebenza ngempumelelo nenembayo.
shiya impendulo