Sunadaran halittu ne na zamantakewa. Su kuma hawainiya. Dangane da buƙatun tantanin halitta, suna canzawa cikin sauri cikin tsari kuma suna kama wasu kwayoyin halitta a cikin rikitaccen rawa.

Ba gidan wasan kwaikwayo na abincin dare ba. Maimakon haka, waɗannan haɗin gwiwar su ne tushen hanyoyin nazarin halittu. Wasu suna kunna ko kashe kwayoyin halitta. Wasu suna ƙwanƙwasa ƙwayoyin “zombie” tsufa don lalata kansu ko kiyaye fahimtarmu da ƙwaƙwalwarmu cikin siffa ta sama ta hanyar sake fasalin hanyoyin sadarwar kwakwalwa.

Waɗannan haɗin gwiwar sun riga sun yi wahayi zuwa ga nau'ikan hanyoyin kwantar da hankali-kuma sabbin hanyoyin warkewa na iya haɓaka ta AI waɗanda zasu iya ƙira da ƙirar ƙwayoyin halitta. Amma kayan aikin AI da suka gabata sun mayar da hankali ne kawai ga sunadaran da hulɗar su, suna watsar da abokan aikinsu marasa gina jiki.

Wannan makon, binciken in Science fadada ikon AI don yin samfura iri-iri iri-iri na sauran kwayoyin halitta waɗanda ke kama jikin sunadaran jiki, gami da ƙananan ƙwayoyin da ke ɗauke da baƙin ƙarfe waɗanda ke zama cibiyar masu ɗaukar iskar oxygen.

Dokta David Baker ya jagoranta a Jami'ar Washington, sabon AI yana faɗaɗa fa'idar ƙirar biomolecular. Wanda ake yiwa lakabi da RoseTTAFold All-Atom, yana ginawa akan tsarin furotin-kawai da ya gabata don haɗa ɗimbin sauran ƙwayoyin halitta, kamar DNA da RNA. Hakanan yana ƙara ƙananan ƙwayoyin cuta-misali, ƙarfe-wanda ke da alaƙa da wasu ayyukan furotin.

AI ta koya ne kawai daga jeri da tsarin abubuwan da aka haɗa-ba tare da wani ra'ayi na tsarin su na 3D ba-amma yana iya fitar da hadadden injunan ƙwayoyin cuta a matakin atomic.

A cikin binciken, lokacin da aka haɗa tare da AI mai haɓakawa, RoseTTAFold All-Atom ya ƙirƙira sunadaran da ke ɗauka cikin sauƙi a kan maganin cututtukan zuciya. Algorithm din ya kuma haifar da sunadaran da ke daidaita heme, kwayoyin halitta mai arzikin ƙarfe wanda ke taimaka wa jini ɗaukar iskar oxygen, da kuma bilin, wani sinadari a cikin tsire-tsire da ƙwayoyin cuta waɗanda ke ɗaukar haske don metabolism.

Waɗannan misalai ne kawai hujjoji na ra'ayi. Ƙungiyar tana sakin RoseTTAFold All-Atom ga jama'a don masana kimiyya don su iya ƙirƙirar abubuwa masu mu'amala da yawa tare da ƙarin rikitarwa fiye da rukunin furotin kaɗai. Bi da bi, abubuwan da aka halitta na iya haifar da sababbin hanyoyin kwantar da hankali.

"Manufarmu a nan ita ce gina kayan aiki na AI wanda zai iya samar da ƙarin hanyoyin kwantar da hankali da sauran kwayoyin halitta masu amfani," in ji marubucin binciken Woody Ahern a cikin wata sanarwa.

Mafarki A

A cikin 2020, Google DeepMind's AlphaFold da Baker Lab's RoseTTAFold sun warware matsalar tsinkayar tsarin furotin wanda ya ba masana kimiyya mamaki tsawon rabin karni kuma ya haifar da sabon zamanin binciken furotin. Sabbin sigogin waɗannan algorithms sun tsara duk tsarin gina jiki wanda aka sani da wanda kimiyya ba a san shi ba.

Na gaba, AI mai haɓakawa—fasaharar da ke bayan OpenAI's ChatGPT da Google's Gemini-ya haifar da hatsaniya na sunadaran ƙira tare da kewayon ayyuka masu ban sha'awa. Wasu sabbin sunadaran da aka samar sun tsara wani hormone wanda ke kiyaye matakan calcium a cikin bincike. Wasu sun haifar da enzymes na wucin gadi ko sunadaran da zasu iya da sauri canza siffar su kamar transistor a cikin da'irori na lantarki.

Ta hanyar hasashe sabuwar duniya na tsarin gina jiki, AI mai haɓakawa yana da yuwuwar yin mafarkin ƙarni na sunadaran sunadarai don daidaita ilimin halittarmu da lafiyarmu.

Amma akwai matsala. Samfuran sunadaran sunadaran AI suna da hangen nesa: Su ne kuma mayar da hankali ga sunadaran.

Lokacin da ake hango abubuwan da ke tattare da kwayoyin halitta na rayuwa, sunadaran sunadaran, DNA, da fatty acid suna zuwa a zuciya. Amma a cikin tantanin halitta, waɗannan sifofi galibi ana haɗa su tare da ƙananan ƙwayoyin cuta waɗanda ke haɗa abubuwan da ke kewaye da su, tare suna samar da babban taro mai aiki.

Misali daya shine heme, kwayar halitta mai kama da zobe wacce ta hada da karfe. Heme shine tushen haemoglobin a cikin sel jajayen jini, wanda ke jujjuya iskar oxygen a cikin jiki kuma yana kama “ƙugiyoyin” sunadaran da ke kewaye da su ta amfani da nau'ikan sinadarai iri-iri.

Ba kamar sunadaran ko DNA ba, waɗanda za a iya ƙirƙira su azaman silar “wasiƙun” kwayoyin halitta, ƙananan ƙwayoyin cuta da mu’amalarsu suna da wuyar kamawa. Amma suna da mahimmanci ga hadaddun injunan kwayoyin halitta kuma suna iya canza ayyukansu sosai.

Abin da ya sa, a cikin sabon binciken nasu, masu binciken sun yi niyyar faɗaɗa ikon AI fiye da sunadaran.

"Mun tashi don haɓaka hanyar tsinkayar tsarin da ke da ikon samar da haɗin gwiwar 3D ga duk kwayoyin halitta" don kwayoyin halitta, ciki har da sunadarai, DNA, da sauran gyare-gyare, marubutan sun rubuta a cikin takarda.

Kungiyar Tag

Tawagar ta fara ne ta hanyar gyaggyara ƙirar furotin da ta gabata AI don haɗa wasu ƙwayoyin cuta.

AI yana aiki akan matakai uku: Na farko yana nazarin jerin “wasiƙa” mai girma ɗaya na furotin, kamar kalmomi akan shafi. Bayan haka, taswirar 2D tana bin diddigin nisan kowace “kalmar” sunadaran daga wata. A ƙarshe, daidaitawar 3D-kamar GPS-taswirar tsarin tsarin furotin.

Sai haɓakawa ya zo. Don haɗa ƙananan bayanan ƙwayoyin cuta a cikin ƙirar, ƙungiyar ta ƙara bayanai game da rukunin atomic da haɗin sinadarai a cikin yadudduka biyu na farko.

A na uku, sun mai da hankali kan chirality - wato, idan tsarin sinadari yana hagu ko dama. Kamar hannayenmu, sinadarai kuma na iya samun tsarin madubi da su sakamakon nazarin halittu masu bambanta. Kamar sanya safar hannu, daidaitaccen “hannu” na sinadari ne kawai zai iya dacewa da “hannun hannu” da aka bayar.

An horar da RoseTTAFold All-Atom akan ma'ajin bayanai da yawa tare da dubban daruruwan bayanan bayanai da ke kwatanta sunadarai, ƙananan ƙwayoyin cuta, da hulɗar su. Daga ƙarshe, ta koyi ƙayyadaddun ƙayyadaddun ƙayyadaddun ƙwayoyin ƙwayoyin cuta masu amfani don gina majalissar furotin. A matsayin duban tsafta, ƙungiyar ta kuma ƙara "ma'aunin amincewa" don gano tsinkaya mai inganci-waɗanda ke haifar da kwanciyar hankali da taruwa masu aiki.

Ba kamar nau'ikan furotin-kawai AI na baya ba, RoseTTAFold All-Atom “na iya yin ƙirar cikakken tsarin biomolecular,” ƙungiyar ta rubuta.

A cikin jerin gwaje-gwajen, ƙirar da aka haɓaka ya fi hanyoyin da suka gabata lokacin da ake koyon "dock" ƙananan ƙwayoyin cuta a kan wani furotin da aka ba da shi-wani mahimmin ɓangaren gano magunguna-ta hanzarin tsinkayar hulɗar tsakanin sunadarai da kwayoyin da ba na gina jiki ba.

Brave New World

Haɗa ƙananan ƙwayoyin cuta yana buɗe sabon matakin ƙirar furotin na al'ada.

A matsayin hujja na ra'ayi, ƙungiyar ta lalata RoseTTAFold All-Atom tare da ƙirar AI ta haɓaka da suke da ita baya ci gaba da kuma tsara abokan haɗin gwiwar furotin don ƙananan ƙwayoyin cuta daban-daban guda uku.

Na farko shine digoxigenin, wanda ake amfani dashi don magance cututtukan zuciya amma yana iya samun illa. Protein da ke kama shi yana rage yawan guba. Ko da ba tare da sanin ƙwayar ƙwayar cuta ba, AI ta tsara nau'ikan furotin da yawa waɗanda ke damun matakan digoxigenin lokacin da aka gwada su a cikin ƙwayoyin halitta.

AI ta kuma tsara sunadaran da ke ɗaure da heme, ƙaramin ƙwayar ƙwayar cuta mai mahimmanci don isar da iskar oxygen a cikin jajayen ƙwayoyin jini, da kuma bilin, wanda ke taimakawa nau'ikan halittun su sami haske.

Ba kamar hanyoyin da suka gabata ba, ƙungiyar ta bayyana, AI na iya "samar da sunadaran sunadaran a hankali" waɗanda ke kama kan ƙananan ƙwayoyin cuta ba tare da sanin ƙwararru ba.

Hakanan yana iya yin tsinkaya mai inganci sosai game da ƙarfin haɗin kai tsakanin sunadarai da ƙananan ƙwayoyin cuta a matakin atomic, yana ba da damar gina sabuwar sararin samaniya ta hanyar hankali.

Baker ya ce "Ta hanyar ƙarfafa masana kimiyya a ko'ina don samar da kwayoyin halitta tare da daidaitattun da ba a taɓa gani ba, muna buɗe kofa ga bincike mai zurfi da aikace-aikace masu amfani waɗanda za su tsara makomar magani, kimiyyar kayan aiki, da kuma bayanta," in ji Baker.

Kirkirar Hoto: Ian C. Haydon

• Ƙunshin Ƙarfafa SEO & Rarraba PR. Samun Girma a Yau.
• PlatoData.Network Tsaye Generative Ai. Karfafa Kanka. Shiga Nan.
• PlatoAiStream. Web3 Mai hankali. Ilmi Ya Faru. Shiga Nan.
• PlatoESG. Carbon, CleanTech, Kuzari, Muhalli, Solar, Gudanar da Sharar gida. Shiga Nan.
• PlatoHealth. Biotech da Clinical Trials Intelligence. Shiga Nan.
• Source: https://singularityhub.com/2024/03/08/this-ai-can-design-biomolecular-machines-with-atomic-precision/