Raka'a
1. Gabatarwa zuwa Hankali na Artificial a cikin Halittar Halitta da Halitta: Matsayi, Iyakoki, Tabbatarwa, Da'a da Keɓantawa 2. Binciken Jeri na DNA/RNA: Daidaitawa, Motif, Buɗe Tsarin Karatu da Basic Bioinformatics 3. Fassarar Bambance-bambance: Halittu ta VCF, Tsarin HGVS, da Ma'auni na ACMG 4. Tsarin Protein da AlphaFold: Hasashen Tsarin Tsarin Karatu, Makin Amincewa, da Tabbatarwa 5. Taimakon ƙira na CRISPR: Zaɓin gRNA, Tasirin Kashe Manufa, da Tabbatar da Gwaji 6. Binciken Bayanan Omics: RNA-seq, Bayyanawa, Ƙididdiga da Haɓaka Tafarki 7. Bita na Adabi da Rubutun Kimiyya: Tabbatar da Tushen da Hatsarin Maganar Ƙarya 8. Fassarar asibiti: Rahoto, Amincewar Kwararru, Tabbatarwa, da Iyaka 9. Ladabi na Hallucination da Tabbatarwa: Ƙwayoyin Halitta, Jeri, Bambance-bambance, da Halaye 10. Da'a, Sirri da Kariyar Bayanai: Alhakin Musamman na Bayanan Halitta 11. Shari'ar Ƙarshe-zuwa-Ƙarshe: Tafiyar Bambancin Daga Ganowa zuwa Rahoton Clinical
Naúrar 11 / 11

Shari'ar Ƙarshe-zuwa-Ƙarshe: Tafiyar Bambancin Daga Ganowa zuwa Rahoton Clinical

Riba:

  • Ability don haɗa duk sassan da aka koya a cikin tsarin (jeri, bambance-bambancen, tsari, wallafe-wallafe, na asibiti) a cikin guda ɗaya na tallafi na AI-ƙarshen aiki
  • Ikon yin amfani da ƙa'idar 'hankali na wucin gadi yana samarwa, tabbatar da ɗan adam da takaddun shaida' da tabbatar da tushen asali a kowane mataki.
  • Ikon sarrafa sarkar tabbatarwa ta hanyar tattara bayanan sirri da amincewar ƙwararrun ƙwararru na ƙarshe a cikin shari'ar ƙarshe zuwa ƙarshe

A cikin wannan tsarin, mun tattauna kowane mataki daban, daga jerin bincike zuwa tsarin gina jiki, daga CRISPR zuwa omics, daga wallafe-wallafe zuwa xa'a da sirri. A cikin wannan naúrar ta ƙarshe, mun haɗa su duka zuwa cikin ingantaccen aiki guda ɗaya: tafiya na bambance-bambancen da aka samu a cikin majiyyaci, daga ɗanyen bayanai zuwa rahoton asibiti da aka yarda. Manufar ita ce ganin cikakken yadda za a sanya hankali na wucin gadi (AI) a matsayin mataimaki na ƙarshe zuwa ƙarshe da kuma yadda tabbatarwa da lissafin ɗan adam ke haɗuwa a kowane mataki.

Labarin mu (na almara): Wata mace mai haƙuri mai shekaru 34 ta zo don kimanta kwayoyin halitta tare da tarihin iyali na ciwon daji na nono tun tana ƙarami. An sami bambance-bambance a cikin kwayar halittar BRCA1 a cikin gwajin kwamitin. Manufar mu: rarraba wannan bambance-bambancen, fassara tasirinsa akan furotin, duba wallafe-wallafen, da kuma rubuta rahoton asibiti - ta amfani da AI a kowane mataki da kuma tabbatar da shi a kowane mataki.

Mataki na 1: Gyara bambance-bambancen da kulawar sirri

A farkon farkon aikin, muna yin abubuwa biyu: daidaita bayanai da tabbatar da sirri.

Sirri na farko: Sunan majiyyaci, lambar tantancewa, da ranar haihuwa ba sa shigar da buɗaɗɗen kayan aikin AI (Raka'a 10). Muna aiki ne kawai tare da bambance-bambancen, sigar genome, da kwafi - kuma kawai idan bambance-bambancen bai gano mutum ba. Sannan mu gyara bambance-bambancen: GRCh38, BRCA1, NM_007294.4, misali c.5266dupC (p.Gln1756fs). Mun tabbatar da wannan tare da mai inganci (VariantValidator); Ba mu yarda da wakilcin da AI ya bayar a ƙimar fuska ba (Raka'a 1, 3).

Mataki na 2: Tara shaida (Sashe na 3)

Muna da AI ta gina tsarin hujja na ACMG - amma ba tare da sanya aji ba. Sannan mu tabbatar da kowace hujja da kanmu:

  • Mitar gnomAD: Muna neman bambance-bambancen a cikin gnomAD kuma mu ga cewa yana da wuya sosai / babu (a cikin shugabanci na PM2).
  • ClinVar: Muna buɗe rikodin kuma bincika maganganun da ke akwai da sabo.
  • Nau'in Bambanci: Idan firamshift ya haifar da dakatarwa da wuri, za a iya samun ƙwaƙƙwaran shaidar asarar aiki (PVS1) - amma muna kimanta wannan a cikin mahallin kwayar halitta da kwafi.

Ba ma amfani da kowane mita ko tunani da AI ke bayarwa ba tare da tabbatar da shi ba (Raka'a 9).

Mataki na 3: Fassarar-Tsarin Protein (Sashe na 4)

Don fahimtar tasirin bambance-bambancen akan furotin, muna zazzage tsarin AlphaFold (tare da lambar UniProt), duba ƙimar amincewar pLDDT da bayanan aikin UniProt na yankin da abin ya shafa. Saboda firamshift yana canza yawancin furotin, fassarar tsarin na iya tallafawa hasashe hasashe a nan-amma tsarin hasashe ne, ba tabbataccen hujja ba. Ba ma tambayar AI don daidaitawa; Mun karanta daga fayil ɗin.

Mataki na 4: Tabbatar da adabi (Sashe na 7)

Muna samun sharuɗɗan bincike daga AI, bincika shi akan PubMed kanta, kuma mu sami ainihin labarai akan bambance-bambancen da kwayar halitta. Muna tabbatar da abubuwan da AI ta gabatar "daga ƙwaƙwalwar ajiya" tare da DOI / PMID. Muna adana bambance-bambance tsakanin dangantaka da dalili a tushen.

Mataki na 5: Rabewa da fassarar asibiti (Raka'a 3, 8)

Mun haɗu da ingantattun shaida, ƙaddamar da nau'i mai yiwuwa tare da haɗin ACMG (a cikin wannan misali, asarar aiki mai karfi da kuma shaida maras kyau na iya kasancewa a cikin hanyar "pathogenic / mai yiwuwa pathogenic" - amma gwani yana da magana ta ƙarshe). Mun haɗu da fassarar asibiti tare da phenotype da tarihin iyali: shin hoton majiyyaci ya yi daidai da BRCA1? Idan akwai rashin tabbas, mun bayyana shi da gaskiya.

Mataki na 6: Rahoton daftarin aiki da amincewa (Sashe na 8)

Tare da AI, muna shirya daftarin rahotanni ga likita da mai haƙuri; A cikin harshen da ba ya wuce gona da iri, bayyananne game da rashin tabbas, da kuma kai tsaye ga shawarwarin kwayoyin halitta. Masanin ya tabbatar da kowace jumla. Muna ƙara sashin iyaka (abin da gwajin bai rufe ba, yanayin sake dubawa).

Tukwici: Lokacin shigar da fitarwa na kowane mataki zuwa na gaba a cikin kwararar ƙarshe zuwa ƙarshen, tabbatar da ingancin fitarwa. Ana aiwatar da mitar da ba a tantance ba ko sifa ta gaba ba daidai ba zuwa duk matakai na gaba ( sarkar hallucination — Unit 9).

tebur mai gudana

Mataki

Matsayin AI

Tabbaci na wajibi

mai yanke shawara

gyarawa

demo zane

VariantValidator

gwani

Keɓantawa

Ikon ɓoyewa

sarrafa mutane

gwani

shaida

Saita tsari

GnomAD, ClinVar

gwani

tsari

Sharhi

pLDDT, fayil

gwani

adabi

Kalmar bincike, taƙaice

DOI/PMID

gwani

aji

hade daftarin

ACMG dokokin

gwani

Rahoton

daftarin harshe

hukuncin amincewa da jumla

Kwararre/Mashawara

Karamin kararraki guda uku (matsaloli a cikin kwarara)

Case 1 - Rubutun da ba daidai ba. Ƙungiya ɗaya ta yi aiki tare da rubutun da AI ta bayar a farkon rafi; daga baya ya gane cewa kwamitin ya yi amfani da wani rubutun daban. c. Matsayinsu ya canza. An warware kuskuren ta hanyar gyarawa daga farkon tare da VariantValidator. Darasi: Kuskure a lokacin gyare-gyare yana rushe dukkan kwararar.

Case 2 - Labari mai goyan bayan karya. A cikin lokacin wallafe-wallafe, AI ya ba da shawarar takarda mai gamsarwa da ke tallafawa rarrabuwa. Binciken PMID ya nuna cewa labarin ba ya wanzu; cire daga jerin shaidu. Haƙiƙanin shaidar ta isa, amma amana na ilimi/na asibiti da ta lalace idan sifa ta ƙarya ta kasance a cikin rahoton.

Case 3 - Tabbatar da samu. A matakin bayar da rahoto, daftarin AI ya haɗa da bayanin "wannan bambance-bambancen tabbas yana haifar da ciwon daji." Mashawarcin ya gyara wannan da cewa "wannan bambance-bambancen yana ƙara haɗarin kamuwa da cutar kansar nono/ovarian, amma baya nuna takamaiman cuta" kuma ya ƙara ba da shawara don shawarwarin kwayoyin halitta. An tsara sautin sauti da daidaito a hannun kwararru.

Samfura guda biyu masu kwafi (ƙarshe zuwa ƙarshe)

1) Mai tsara kwarara:

Matsayinku: mataimaki na aikin jinya na asibiti. Ƙirƙirar jerin abubuwan bincike na ƙarshe zuwa ƙarshe daga bincike don bayar da rahoto don bambance-bambancen masu zuwa: [Siffar kwayoyin halitta, kwayar halitta, kwafin, HGVS]. Rubuta a kowane mataki (gyara, sirri, shaida, tsari, wallafe-wallafe, aji, rahoto) tare da ginshiƙai "Menene AI ​​yi / menene na tabbatar / wanda ya yanke shawara". KA yanke shawarar aji da na asibiti.

2) Dubawa ta ƙarshe:

Kafin a duba cikakken bayanin bambance-bambancen da ke biyo baya da daftarin rahoto kafin ƙaddamarwa: [abun ciki]. Bincika don: mitar da ba a tabbatar da shi ba/sanyi, ƙari mai yawa sanadi/maganin madaidaici, ɓacewar faɗakarwa, rabewar iyaka, rashin lafiyar sirri. Jera kowace matsala da kowane shawarwari don gyara ta.

Rauni mai ƙarfi / Ƙarfi mai ƙarfi

Rauni: "Rubuta cikakken rahoton asibiti don wannan bambance-bambancen BRCA1."

Matsala: AI yana samar da "rahoton" ta hanyar dacewa da duk matakai a cikin tafiya ɗaya; mita, daftarin aiki, aji, hukunci na asibiti duk ba a tabbatar da su ba.

Ƙarfafa: "Yi lissafin bincike daga ganowa don bayar da rahoto ga wannan bambance-bambancen BRCA1; ware a kowane mataki abin da za ku yi, abin da zan tabbatar, da kuma wanda zai yanke shawara. Zan yi aji da yanke shawara na asibiti bayan tabbatar da shaida."

Dalilin da ya sa yake da ƙarfi: An tsara kwararar ruwa, kowane mataki yana da alaƙa da tabbatarwa, kuma yanke shawara da alhakin sun kasance a kan ɗan adam.

Kuskuren gama gari

  • Neman "rahoton gamawa" tare da faɗakarwa guda ɗaya. An haɗa matakan ba tare da tabbatarwa ba; kuskure yana tarawa.
  • Yin watsi da lokacin gyarawa. Siga/rubutun da ba daidai ba ya karyata dukkan rafi.
  • Matsar da fitarwa tsakanin matakai ba tare da inganta shi ba. Sarkar hallucination ya taso.
  • Yin tunani game da keɓantawa a "ƙarshen" ya kwarara. Dole ne a tabbatar da sirri tun farko, a shigar da bayanai.
  • Barin yanke shawara da sautin zuwa AI. Dole ne ƙwararren ya amince da saƙon aji da na asibiti.
Tsanaki: Yin amfani da ƙarshen-zuwa-ƙarshen AI yana adana lokaci mai yawa amma baya rage alhakin; Akasin haka, yana sa tsarin tabbatarwa a kowane mataki ya fi mahimmanci. AI yana hanzarta kwarara; Dan Adam yana bada garantin daidaito, tsaro da bin ɗabi'a.

Zurfin: tanadin lokacin tafiyar aiki, rashin ƙarfi da hanyar duba

Haƙiƙanin ƙimar ƙarshen-zuwa-ƙarshen kwarara ba shine "bari AI ta yi komai ba" amma a cikin samun lokaci a wuraren da suka dace da rage gudu a wuraren da suka dace. Bari mu ƙididdige ribar da aka samu a cikin wannan yanayin: kafa tsarin shaida, samar da sharuddan bincike, sauƙaƙa harshen rahoton, da kuma fito da jerin abubuwan dubawa na iya yanke abin da al'ada zai ɗauki sa'o'i da yawa na tsara aikin zuwa dubun mintuna. Sabanin haka, ingantaccen gnomAD/ClinVar, VariantValidator pinning, DOI dubawa, da ƙwararrun ƙwararrun ba za a iya taƙaita su ba - su ne ƙashin bayan tsaro na rafi. Daidaitaccen tunani: "AI yana samar da daftarin aiki, ayyukan ɗan adam a matsayin garanti." Riba yana cikin daftarin, alhakin yana cikin yanke shawara.

Mafi raunin hanyar haɗin gwiwa a cikin kwarara shine koyaushe na farko: anga. Sigar genome da ba daidai ba ko kwafin da ba daidai ba yana yaduwa zuwa kowane mataki na gaba, a ƙarshe yana samar da rahoton mara inganci. Don haka kar a ci gaba zuwa mataki na gaba ba tare da tabbatar da anga tare da kayan aiki guda ɗaya ba (kamar VariantValidator). Batu na biyu mafi rauni shine abin da ake fitarwa tsakanin matakai: sanya fitowar da ba a tantance ba ta mataki daya shigar da zuwa na gaba yana gayyatar sarkar hasashe.

A ƙarshe, a cikin mahallin asibiti, hanyar duba tana da alaƙa da kwarara: dole ne a rubuta asalin tushen da aka bincika akan wane kwanan wata, wanda aka ga sigar ClinVar, wanda ya amince da wace shawara. ClinVar sake dubawa yana canzawa akan lokaci; Rikodin da ke "mai cutarwa a yau" na iya komawa zuwa VUS bayan shekara guda. Hanya na duba yana ba da damar sake kimantawa da lissafin shari'a/da'a.

yanayin yawo

ka'ida

Aikace-aikace

ceto lokaci

Tashi don gudun kan daftarin

Tsarin, harshe, jerin abubuwan dubawa daga AI

Rashin ƙarfi

Tsare zoben farko

Tabbatar da gyarawa tare da kayan aiki mai zaman kansa

Kayan sufuri

Matsar ba tare da tabbatarwa ba

An yarda da kowane mataki a tushen

hanyar tantancewa

Rubuta kowane yanke shawara

Tushen, kwanan wata, an yarda da rajista

A takaice

  • Tafiyar bambance-bambancen daga ganowa zuwa rahoto; Ya ƙunshi matakan gyarawa, sirri, shaida, tsari, adabi, aji da rahoto.
  • AI yana samar da daftarin aiki / lambar / taƙaitawa a kowane mataki; A kowane mataki, gwani ya tabbatar da yanke shawara.
  • Ya kamata a matsar da fitarwa tsakanin matakai zuwa na gaba kawai bayan tabbatarwa.
  • An tabbatar da sirri tun daga farko; A ƙarshe an yanke shawarar aji da na asibiti tare da amincewar ƙwararru.

Aikin aikace-aikace

Zaɓi bambance-bambancen almara (misali sanannen rikodin BRCA1). 1. Ƙirƙirar jerin abubuwan dubawa na ƙarshe zuwa ƙarshe tare da samfuri kuma aiwatar da kowane mataki da kanka a cikin hanyar naúrar da ta dace a cikin wannan ƙirar: gyara bambance-bambancen, duba gnomAD/ClinVar, duba tsarin AlphaFold, nemo ainihin labarin, tabbatar da aji mai yiwuwa da zayyana sakin layi na rahoto kuma a sake karanta shi tare da ƙwararrun ido. Yi ƙoƙarin nemo bambanci tsakanin AI da tushen aƙalla matakai biyu.

jerin abubuwan dubawa

  • [ ] Na kafa sirri tun daga farko; Ban canja wurin kowane bayanan ganowa ba.
  • [ ] Na gyara bambance-bambancen tare da sigar/rubutu kuma na tabbatar da shi tare da mai inganci.
  • [] Na tabbatar da kowace shaida (yawanci, ClinVar, wallafe-wallafe) a cikin asali na farko.
  • [ ] Na yi sharhin tsarin tare da amintaccen maki kuma daga fayil ɗin.
  • [ ] Ni kaina na tabbatar da aji da fassarar asibiti bisa ga shaida.
  • [ ] Na duba sahihanci/madaidaici a cikin yaren rahoton da ƙarin iyakoki.

Gwajin Module

1. Wanne daga cikin waɗannan shine mafi daidaiton matsayi don basirar wucin gadi (musamman babban tsarin harshe) a cikin ilimin halitta da kwayoyin halitta?

  • A) Mataimaki ne wanda ya rubuta code, samar da zayyana da sharhi; Dole ne a tabbatar da kowane al'amari na halitta a tushen farko, fassarar ƙarshe ta ƙwararren ✔
  • B) Yana da amintaccen bayanan kwayoyin halitta; Za a iya rubuta jeri da bambance-bambancen da aka bayar kai tsaye zuwa rahoton.
  • C) Tun da yake yana auna kamar kayan aikin dakin gwaje-gwaje, abubuwan da aka fitar ana daukar su shaidar gwaji.
  • D) Yana da cancanta don kammala yanke shawara na cututtukan cututtuka na asibiti ba tare da yardar ɗan adam ba.

Bayani: Babban samfurin harshe ba bayanan kwayoyin halitta bane ko kayan aikin dakin gwaje-gwaje; Janareta ce ta rubutu wacce ke kwaikwayi nassosi a bayanan horo. Ya kamata a tabbatar da gaskiyar abubuwan halitta kamar jerin / bambance-bambancen / gene a koyaushe daga tushe na farko kamar NCBI, Ensembl, ClinVar, gnomAD; Fassarar asibiti na ƙarshe da na kimiyya yakamata ya kasance na ƙwararren gwani.

2. Kun karɓi tsarin haɗin gwiwar nau'ikan nau'ikan nau'ikan nau'ikan nau'ikan AI, amma aikinku yana amfani da GRCh38 (hg38). Ƙila ilimin wucin gadi ya ba da haɗin gwiwar hg19. Menene hanya madaidaiciya?

  • A) Yin amfani da haɗin kai kai tsaye saboda yana da ma'ana
  • B) Babu buƙatar neman sigar kamar yadda hankali na wucin gadi koyaushe yana ba da mafi kyawun sigar yanzu.
  • C) A sarari tabbatar da sigar genome kuma tabbatar da daidaitawa tare da kayan aikin liftOver na hukuma ✔
  • D) Tun da bambanci tsakanin hg19 da hg38 ba shi da mahimmanci, zaɓi ɗaya kuma ci gaba

Bayani: Yana daidaitawa tsakanin nau'ikan kwayoyin halitta (GRCh37/hg19 da GRCh38/hg38). Duk wani haɗin gwiwar da aka bayar ba tare da siga ba ana tuhuma. Ya kamata a canza / tabbatar da haɗin kai tare da kayan aikin liftOver na hukuma kuma ya kamata a tabbatar da sigar genome da aka yi amfani da ita a fili; in ba haka ba duk daidaitawa da sharhi za su nuna wurin da ba daidai ba.

3. A cikin bayanin HGVS 'c.', 'p.' kuma 'g.' Menene ma'anar prefixes kuma me yasa ba za su ruɗe ba?

  • A) Dukkansu guda ukun suna nufin abu daya ne, ba ruwan wanne ake amfani da su.
  • B) 'c.' codeing DNA/kwafi, 'p.' protein, 'g.' shine matakin genomic; Idan ya rikice, yana iya nuna wani wuri daban ✔
  • C) 'p.' kullum 'c.' ya ninka darajarsa sau uku, jujjuyawar ta atomatik ce kuma mara kuskure
  • D) Prefixes na yau da kullun ne kawai; Babu buƙatar tantance sigar kwafin

Bayani: HGVS shine daidaitaccen tsarin rubutun rubutu tare da bambance-bambancen: 'c.' matakin DNA (kwafin rubutu), 'p.' matakin furotin, 'g.' yana nufin matakin genomic. Haɗa waɗannan ko yin amfani da sigar rubutun da ba daidai ba na iya nuna wani wuri/bambanta gaba ɗaya; saboda haka dole ne a yi amfani da madaidaicin prefix tare da sigar kwafin (misali NM_007294.4).

4. Menene ya kamata ya zama aikin AI yayin amfani da ma'auni na ACMG don tantance cututtukan cututtuka na bambance-bambancen?

  • A) Ƙwararrun Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwararren Ƙwaƙwalwa na Ƙarƙashin Ƙarƙashin Ƙarƙashin Ƙarƙashin Ƙarƙashin Ƙarƙashin Ƙarya da Ƙirar Ƙiƙa ne; babu buƙatar kulawar ƙwararru
  • B) Gudanar da GnomAD ba lallai ba ne tunda hankali na wucin gadi yana ba da mitar yawan jama'a daga ƙwaƙwalwar ajiyar sa.
  • C) Tun da ma'aunin ACMG yana da tsauri, ana ɗaukar fitowar AI ta atomatik daidai
  • D) Hankali na wucin gadi na iya samar da daftarin shaida; Duk wata shaida yakamata a tabbatar da ita a asali na farko, ƙwararrun ƙwararrun ✔

Bayani: ACMG wani tsari ne wanda ke rarraba bambance-bambancen cututtukan cututtuka ta nau'ikan shaida (mitar yawan jama'a, a cikin hasashen siliki, rarrabuwa, nazarin aiki, adabi). AI na iya samar da daftarin shaida; Duk da haka, kowace shaida dole ne a tabbatar da kanta a cikin ClinVar, gnomAD da wallafe-wallafen, kuma ƙwararren ƙwararren ƙwararren ƙwararren ya yi rarrabuwa na ƙarshe (mai yiwuwa pathogenic / VUS ...).

5. Za ku yi aiki tare da jerin DNA da aka bayar ta hanyar basirar wucin gadi. Menene mafi kyawun matakin farko?

  • A) Yin amfani da jerin ta hanyar samun shi daga tushe na farko kamar NCBI / Ensembl ko tabbatar da shi tare da tunani a can ✔
  • B) Shigar da jeri kai tsaye cikin bincike domin ya bayyana ya zama daidai tsayi
  • C) Zaton jerin daidai ne idan haruffan sun ƙunshi ingantattun nucleotides
  • D) Tambayi hankali na wucin gadi game da jerin kuma sake ƙidaya shi daidai idan abubuwan guda biyu iri ɗaya ne.

Bayani: AI na iya 'cika' igiyar da ba ta tunawa da haruffa masu kyan gani; Ko da tsayi da haruffa sun bayyana daidai, ƙila ba za su dace da ainihin abin da ake magana ba. Saboda haka, duk wani jerin da aka bayar ta hanyar basirar wucin gadi daga kai ba a amfani da shi kai tsaye; Ana ɗaukar jerin daga tushe na farko kamar NCBI / Ensembl ko an tabbatar da shi ta hanyar tunani a ciki.

6. Menene ma'aunin pLDDT ke nufi a cikin hasashen tsarin furotin daga kayan aiki kamar AlphaFold?

  • A) Takaddun shaida ce da ke nuna cewa an tabbatar da tsarin da gwaji
  • B) Ƙimar dakin gwaje-gwaje ce da ke auna matakin aiki na furotin.
  • C) Maki ne da ke nuna amincewar samfurin a cikin hasashen tsarin gida na kowane saura; ƙananan ƙima yana sassauƙa / yanki mara dogaro ✔
  • D) Rabo ne da ke nuna bambancin mitar furotin a cikin jama'a.

Bayani: pLDDT (0-100) ƙimar amincewa ce wacce ke nuna yadda ƙarfin AlphaFold ke cikin hasashen tsarin sa na gida don kowane ragowar (amino acid). Babban pLDDT yana nufin ingantaccen tsarin gida, ƙananan pLDDT yana nufin sassauƙa / rashin tsari ko yanki mara dogaro. Tsarin shine zato; Yankunan da ba su da ƙarfin gwiwa kada su kasance masu fassarori da yawa kuma ya kamata a goyi bayan da'awar aiki ta hanyar shaidar gwaji.

7. Me yasa yake da haɗari a yi iƙirarin cewa maye gurbi 'yana hana' sunadaran da ya dogara akan tsarin Hasashen AlphaFold?

  • A) Ba za a iya amfani da tsarin AlphaFold don kowane dalili ba saboda koyaushe kuskure ne
  • B) Tsarin zato ne; da'awar aiki suna a matakin hasashe kuma suna buƙatar tabbatarwa na gwaji ✔
  • C) Babu haɗari; Fitowar AlphaFold daidai yake kamar ginin gwaji
  • D) Akwai haɗari kawai idan pLDDT yayi ƙasa; Idan yana da girma, ana ɗaukar da'awar ta atomatik.

Bayyanawa: AlphaFold hasashe ne na tsari, ba hujjar gwaji ba; Yana da iyakance musamman a yankuna masu sassauƙa / rashin tsari kuma a cikin ƙirar tasirin maye gurbi. Da'awar aiki bisa ginin suna a matakin hasashe kuma suna buƙatar tabbatarwa na gwaji (misali, nazarin aiki). In ba haka ba, an gabatar da hasashen a matsayin shaida.

8. Menene mafi mahimmancin matakin tabbatarwa game da CRISPR gRNA (jagorancin RNA) 'yan takarar da aka ba da shawara ta hanyar basirar wucin gadi?

  • A) Idan tsayin gRNA daidai ne, ba a buƙatar ƙarin bincike
  • B) Idan basirar wucin gadi ta ce 'tasirin kashe-kashe ba shi da yawa', ci gaba kai tsaye zuwa gwajin
  • C) Ya isa tabbacin cewa jerin gRNA ya ƙunshi ingantattun nucleotides
  • D) Tabbatar da 'yan takara tare da daidaitawar kwayoyin halitta na gaskiya da ƙwararrun kayan aikin bincike-bincike kuma tabbatarwa a cikin dakin gwaje-gwaje ✔

Bayani: gRNA shine jagorar RNA wanda ke jagorantar enzyme Cas zuwa manufa; Tare da ƙirar da ba daidai ba, yana iya kuma yanke wuraren da aka yi niyya. Dole ne 'yan takarar da AI suka dawo da su ta hanyar ainihin daidaitawar kwayoyin halitta da ƙwararrun kayan aikin bincike-bincike, sannan an tabbatar da su ta hanyar gwajin dakin gwaje-gwaje. Kawai saboda hankali na wucin gadi ya ce 'lafiya' ba shine madadin tabbatarwa ba.

9. Me yasa gyaran gwaji da yawa (misali FDR/Benjamini-Hochberg) ya zama dole yayin gwada bambance-bambance a cikin maganganun dubban kwayoyin halitta tare da RNA-seq?

  • A) Dubban gwaje-gwaje na lokaci guda suna haifar da halayen karya; Gyaran FDR yana iyakance waɗannan, ɗanyen p-darajar yaudara ce ✔
  • B) Gyara ya zama dole ne kawai idan adadin samfurori kadan ne; Ba shi da alaƙa da adadin kwayoyin halitta
  • C) Gyaran gwaji da yawa baya canza sakamakon, yana da tsari ne kawai
  • D) Danyen p<0.05 koyaushe ya isa; gyara yana sa ƙididdiga ta kasance da wahala ba dole ba

Bayani: Lokacin da aka gwada dubban kwayoyin halitta a lokaci guda, yawancin kwayoyin halitta suna zuwa da p<0.05 (ƙarya mai kyau) saboda dama kadai. Gyaran gwaji da yawa (FDR - sarrafa ƙimar gano ƙarya) yana iyakance waɗannan halayen karya. Jerin kwayoyin halittar da aka ayyana 'mahimmanci' tare da p-darajar da ba a gyara ba suna yaudara; Ya kamata a yi amfani da ƙimar da aka daidaita (q-darajar).

10. Menene ma'anar idan hanya aka gano tana da 'inganta sosai' a cikin nazarin haɓakar hanyar kuma ta yaya za a fassara ta?

  • A) Ya tabbatar da cewa hanya ita ce tabbatacciyar dalilin cutar
  • B) Shaida ce ta gwaji cewa hanyar tana aiki a cikin tantanin halitta
  • C) Siginar ƙididdiga ce da ke nuni da cewa ƙwayoyin halittar hanya sun yi yawa fiye da yadda ake tsammani; Yana buƙatar fassarar mahimmanci, ba dalili ba ✔
  • D) Isasshen sakamako ne a kan kansa, mai zaman kansa na asali da gyaran gwaji da yawa

Bayani: Haɓaka sigina ce ta ƙididdiga da ke nuna cewa wata hanya ta musamman ta cika da yawa a tsakanin kwayoyin halitta waɗanda bayyanarsu ta canza; Ba hujja ba ce ta sanadin ko kuma a zahiri hanyar tana aiki. Ƙarshe; Dole ne a fassara saitin kwayoyin halitta da aka yi amfani da shi da mahimmanci a cikin mahallin zaɓi na baya da gyaran gwaji da yawa tare da ƙwarewar ilimin halitta.

11. Lokacin ba da taƙaitaccen wallafe-wallafe, AI na iya buga labarin da ya bayyana na gaske amma babu shi (ƙirar karya). Menene hanya madaidaiciya?

  • A) Idan marubucin da sunan jarida na gaskiya ne, ƙila ambaton ya yi daidai, bincika ba lallai ba ne
  • B) Tabbatar da kowace magana a cikin tushen farko ta DOI/PubMed; ✔ rashin amfani da abin da ba a iya samu ba
  • C) Hankali na wucin gadi ba ya haɗa da sifa; Za a iya ƙara nassoshi kai tsaye zuwa littafin littafin
  • D) Idan nassoshi suna da yawa, ba matsala idan kaɗan daga cikinsu sun yi kuskure.

Bayani: AI na iya samar da labarin da aka yi gaba ɗaya (da DOI na karya) tare da sunaye masu sauti na ainihi a cikin mujallu na gaske. Dole ne a tabbatar da kowace magana da kanta a tushen farko ta DOI/PubMed; Kada a yi amfani da nassoshi waɗanda ba za a iya samun su ba. Alhakin kimiyya na marubuci ne; Babu wasu bayanan da ba a tantance ba da ya kamata su shigar da rubutun.

12. Menene madaidaicin ƙa'ida game da zayyana rahoton rahoton cututtukan asibiti da AI ya haifar?

  • A) Idan daftarin yana da kyau, ana iya ba shi kai tsaye ga majiyyaci
  • B) Ba a buƙatar tabbatar da tushe daban kamar yadda AI kuma ke haifar da shaida
  • C) Za a iya barin yanke shawara mai cutarwa a cikin rahoton zuwa hankali na wucin gadi, gwani kawai yana duba tsarin
  • D) Kowane da'awar asibiti dole ne a tabbatar da shi, rahoton ƙarshe dole ne a amince da ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun masu ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun masu fafutuka, da kuma ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun ƙwararrun qungiyoyi, da suka amince da bayar da rahoton qarshe, kuma dole ne a rubuta jerin sassan tabbatarwa ✔

Bayani: Leken asiri na wucin gadi yana adana lokaci wajen samar da daftarin rahoto da taƙaitawa; Duk da haka, kowane da'awar asibiti dole ne a haɗa shi da shaida da tushe na farko, kuma ƙwararren mai izini dole ne ya amince da rahoton ƙarshe. Ba za a iya fassara fitowar AI kai tsaye zuwa yanke shawara mai haƙuri ba; Dole ne a rubuta sarkar tabbatarwa.

13. Menene ma'anar idan fitowar AI ta yi sauti sosai kuma 'm'?

  • A) Sautin amintacce ba shaida ba ce ta daidaito; Bukatar tabbatarwa yana ƙaruwa, ba raguwa ba ✔
  • B) Amsa mai kyau da yanke hukunci mai yuwuwa gaskiya ce, tabbatarwa ba lallai ba ne
  • C) Sautin aminci yana nuna samfurin yana karanta bayanai kai tsaye daga ƙwaƙwalwar ajiya
  • D) Tone shine ma'auni na amincewa, saboda amsoshi masu shakka ba daidai ba ne kuma amsoshi masu tabbaci suna daidai.

Bayani: Amintaccen sautin AI ba shaida ba ne na daidaito; Mafi hatsarin hasashe sun zo tare da mafi kyawun magana da gamsarwa. Lokacin da aka ga sautin tabbaci, buƙatar tabbatarwa yana ƙaruwa, ba raguwa ba. Kowace hujja (jeri, bambance-bambancen, jinsi, tunani) yakamata a inganta shi tare da tushen asali mai zaman kansa.

14. Menene mafi mahimmancin kariya ta sirri kafin ba da bayanan kwayoyin halitta na majiyyaci zuwa kayan aikin AI na jama'a don bincike?

  • A) Idan bayanan an sanya su a cikin base64, ana iya manna su lafiya
  • B) Idan kawai an goge sunan majiyyaci, ba a buƙatar ƙarin wani mataki
  • C) Rashin samar da bayanan da za a iya gane su kwata-kwata; ɓata suna, rage yawan bayanai da yarda / kiyaye tsarin doka ✔
  • D) Ana iya raba bayanan kwayoyin halitta kyauta saboda ba a la'akari da bayanan sirri ba.

Bayani: Bayanan Halitta wani nau'i ne na musamman na bayanan sirri kuma ana iya sake gano mutum tare da haɗakar bambance-bambancen da ba kasafai ba. Ba a taɓa ba da bayanan kwayoyin da za a iya ganewa don buɗe kayan aikin ba; ya kamata a ɓoye bayanan, kawai mafi ƙarancin bayanan da ake buƙata ya kamata a raba (rage girman bayanai) kuma a kiyaye yarda/tsarin doka (KVKK/GDPR).

A baya Kammala tsarin