Ma `lumot

Hujayralarning o'sishini modellashtirish va vizualizatsiya qilish vositalari

Hujayralarning o'sishini modellashtirish va vizualizatsiya qilish vositalari


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Men hujayra o'sishi uchun simulyatsiya qilishda yordam beradigan dastur/GUI ilovasi/paketini qidiryapman. Menda mikroskopik matematik model bor va stsenariy asosan quyidagicha: men uning markazi va yo'nalishi burchagini tavsiflovchi novda shaklidagi hujayradan boshlayman, u ma'lum uzunlikka yetguncha o'sadi, burchaklari tasodifiy o'zgarib turadigan ikkita bir xil hujayraga bo'linadi. va jarayon davom etmoqda.

Bunday vaziyatda foydalanishim mumkin bo'lgan biron bir dastur yoki paket bormi? Men GitHub-da bir nechtasini topdim, lekin ular asosan juda murakkab biologik jarayonlar uchun edi. Men Python/Julia/Matlab-dan foydalanaman va kodlashim mumkin, lekin men bunday oddiy stsenariylar uchun vositalar bormi, deb qiziqdim.


Biologik tizimlarni modellashtirish

Biologik tizimlarni modellashtirish tizim biologiyasi va matematik biologiyaning muhim vazifasidir. [a] Hisoblash tizimlari biologiyasi [b] [1] biologik tizimlarni kompyuterda modellashtirish maqsadida samarali algoritmlar, ma'lumotlar tuzilmalari, vizualizatsiya va aloqa vositalarini ishlab chiqish va ishlatishga qaratilgan. Bu hujayra jarayonlarining murakkab aloqalarini tahlil qilish va vizualizatsiya qilish uchun biologik tizimlarning kompyuter simulyatsiyasidan, shu jumladan uyali quyi tizimlardan (metabolizm, signal uzatish yo'llari va genlarni tartibga solish tarmoqlarini o'z ichiga olgan metabolitlar va fermentlar tarmoqlari) foydalanishni o'z ichiga oladi. [2]

Murakkab tizimning kutilmagan paydo bo'ladigan xususiyati oddiyroq, yaxlit qismlar o'rtasidagi sabab va ta'sirning o'zaro ta'siri natijasi bo'lishi mumkin (qarang: Biologik tashkilot). Biologik tizimlar komponentlarning murakkab o'zaro ta'sirida paydo bo'ladigan xususiyatlarning ko'plab muhim misollarini namoyon qiladi. Biologik tizimlarni an'anaviy o'rganish reduktiv usullarni talab qiladi, bunda ma'lumotlar miqdori toifalar bo'yicha to'planadi, masalan, ma'lum bir stimulga javoban vaqt o'tishi bilan konsentratsiya. Kompyuterlar ushbu ma'lumotlarni tahlil qilish va modellashtirish uchun juda muhimdir. Maqsad tizimning atrof-muhit va ichki ogohlantirishlarga reaktsiyasining real vaqt rejimida aniq modellarini yaratishdir, masalan, saraton hujayrasining signalizatsiya yo'llarining zaif tomonlarini topish uchun modeli yoki kardiomiotsitlarga ta'sirini ko'rish uchun ion kanallari mutatsiyalarini modellashtirish. o'z navbatida, yurak urishining funktsiyasi.


Rivojlanayotgan butun hujayralarni modellashtirish tamoyillari va usullari

Butun hujayrali modellar har bir gen funktsiyasini ifodalash orqali genotipdan fenotipni bashorat qiladi.

Butun hujayrali modellar biofan, bioinjeneriya va tibbiyotni o'zgartirishi mumkin.

Butun hujayrali modellarga erishish uchun ko'plab qiyinchiliklar mavjud.

Yangi o'lchash va modellashtirish texnologiyalari jadallik bilan butun hujayrani modellashtirish imkonini beradi.

Kengaytiriladigan modellashtirish vositalarini yaratish bo'yicha davom etayotgan sa'y-harakatlar butun hujayrali modellashtirishni tezlashtiradi.

Butun hujayrali hisoblash modellari genomni, har bir molekulyar turning tuzilishi va kontsentratsiyasini, har bir molekulyar o'zaro ta'sirini va hujayradan tashqari muhitni ifodalash orqali genotipdan uyali fenotiplarni bashorat qilishga qaratilgan. Butun hujayrali modellar biologiya, bioinjeneriya va tibbiyotni o'zgartirish uchun katta imkoniyatlarga ega. Biroq, butun hujayrali modellarga erishish uchun ko'plab qiyinchiliklar mavjud. Shunga qaramay, tadqiqotchilar birinchi butun hujayra modellarini yaratish uchun o'lchash texnologiyasi, bioinformatika, ma'lumotlar almashish, qoidalarga asoslangan modellashtirish va ko'p algoritmli simulyatsiya sohasidagi so'nggi yutuqlardan foydalana boshladilar. Biz butun hujayrali modellashtirish vositalarini ishlab chiqish bo'yicha olib borilayotgan sa'y-harakatlar odamlarning hujayralari modellarini o'z ichiga olgan holda, yanada kengroq va aniqroq modellarga imkon berishini kutmoqdamiz.


Natijalar va muhokama

CRISPR skrining tajribalari uchun sifat nazorati o'lchovlari

MAGECK yordamida asosiy genlarni aniqlashdan tashqari, MAGECK-VISPRning asosiy maqsadi turli darajadagi sifat nazorati (QC) o'lchovlarini yig'ishdir (http://www.bioinformatics.babraham.ac.uk/projects/fastqc). Tavsiya etilgan o'lchovlarni (1 -jadval) to'rt toifaga bo'lish mumkin: ketma -ketlik darajasi, o'qish sonining darajasi, namuna darajasi va gen darajasi (2 -rasm).

VISPR ning sifat nazorati (QC) ko'rinishi, MAGECK-VISPR vizualizatsiya tizimi. O'lchovlar GC tarkibini taqsimlashni o'z ichiga oladi (a), o'rtacha asosiy sifat (b), o'rtacha ketma-ketlik sifatining taqsimlanishi (v), nol sonli sgRNKlar soni (d), Jini indeksi (e), umumiy o'qishlar soni va xaritalangan o'qishlar foizi (f), Asosiy komponentlar tahlili (PCA) syujeti (g), normallashtirilgan o'qishlar soni taqsimoti (h, i) va juftlik namunali korrelyatsiyalar (j). Ko'rsatilgan natijalar ESC (a-f) va melanoma ma'lumotlar to'plami (g-j)

QC darajasining ketma-ketligi o'lchovlari boshqa keyingi avlod sekvensiyasi (NGS) tajribalarida bo'lgani kabi ketma-ketlik bilan bog'liq muammolarni aniqlashga qaratilgan. Ikkita o'lchov haqida xabar berilgan: namuna GC tarkibini taqsimlash (2a-rasm) va ketma-ketlikni o'qishning asosiy sifatini taqsimlash (2b, s-rasm). Ideal holda, ketma-ket o'qishlar asosli asosiy sifatlarga ega bo'lishi kerak (o'rtacha qiymat & gt25) va bir xil tajriba namunalari o'xshash GC kontent taqsimotiga ega bo'lishi kerak.

QC o'lchovlarining ikkinchi darajasi MAGECK dan to'plangan sgRNK o'qish soniga asoslanadi. Xom ketma-ketlik o'qishlari birinchi navbatda kutubxonadagi sgRNK ketma-ketliklari bilan taqqoslanadi, hech qanday nomuvofiqlikka yo'l qo'yilmaydi. Shundan so'ng, har bir namuna uchun ketma-ket o'qishlar soni, xaritalangan o'qishlar (va undan xaritalangan o'qishlar foizi), o'qish soni nolga teng bo'lgan sgRNAlar va o'qishlar sonini taqsimlashning Gini indeksi xabar qilinadi (2d-f-rasm). Xaritalangan o'qishlar foizi namuna sifatining yaxshi ko'rsatkichidir va past xaritalash qobiliyati ketma-ketlik xatosi, oligonukleotid sintez xatosi yoki namunaning ifloslanishi bilan bog'liq bo'lishi mumkin. Pastki oqim tahlilining yaxshi statistik kuchi har bir sgRNK uchun etarli o'qishga (afzalroq 300 dan ortiq o'qish) tayanadi, plazmid kutubxonasida nol soni sgRNKlar soni yoki erta vaqt nuqtalari. Iqtisodiyotda daromadlar tengsizligining umumiy o'lchovi bo'lgan Jini indeksi sgRNK o'qish sonining tengligini o'lchashi mumkin [14]. Ijobiy tanlov tajribalaridagi keyingi vaqt nuqtalari Gini indeksining yuqori bo'lishi normal holat, chunki bir nechta omon qolgan klonlar (juda ko'p sonli sgRNK) oxirgi hovuzda hukmronlik qilishi mumkin, boshqa hujayralarning ko'pchiligi o'lsa (nolga teng sGRNA) ). Aksincha, plazmid kutubxonasida, dastlabki vaqt nuqtalarida yoki salbiy tanlov tajribalarida yuqori Jini indeksi mos ravishda CRISPR oligonükleotid sintezining notekisligini, past virusli transfeksiya samaradorligini va ortiqcha tanlovni ko'rsatishi mumkin.

Namuna darajasi QC (2g-j-rasm) namunalar orasidagi moslikni tekshiradi. MAGECK-VISPR normallashtirilgan o'qishlar sonining quti chizmalari va kümülatif taqsimot funktsiyalari bo'yicha taqsimlanishi haqida xabar beradi. Shuningdek, u namunalar jurnalini o'qish sonining juft-juft Pearson korrelyatsiyasini hisoblab chiqadi va namunalarni Principle Component Analysis (PCA) ning dastlabki uchta komponentiga tortadi. Biologik replikatsiyalar yoki shunga o'xshash sharoitga ega namunalar o'xshash o'qishlar soni taqsimotiga va yuqori korrelyatsiyaga ega bo'lishi va PCA uchastkasida bir-biriga yaqinroq ko'rinishi kerak. PCA uchastkalari, agar ekranlar turli partiyalar ostida o'tkazilsa, potentsial partiya effektlarini ham aniqlashi mumkin.

Nihoyat, QC gen darajasi ekranlardagi salbiy tanlov darajasini aniqlaydi. Ribosomal genlarni nokaut qilish kuchli salbiy tanlov fenotipiga olib kelganligi sababli [1, 2], ribosoma genlaridagi salbiy tanlovning ahamiyati MAGECK-VISPR da GOrilla [15] yordamida Gen Ontologiyasi (GO) boyitish tahlili orqali baholanishi mumkin. Ishlaydigan salbiy tanlov tajribasi muhim ahamiyatga ega bo'lishi kerak P qiymat (<0,001), garchi ko'plab yaxshi tajribalar ancha kichikroq bo'lishi mumkin P qiymatlar (<1e-10, Qo'shimcha fayl 1ning A bo'limiga qarang).

MAGECK-MLE bilan bir nechta sharoitlarda muhim genlarni chaqirish

MAGeCK-VISPR o'z ichiga "MAGeCK-MLE" yangi algoritmini o'z ichiga oladi, ehtimollik baholashning (MLE) yondashuvidan foydalanib, har xil skrining sharoitida genlarning muhimligini baholaydi. Faqat ikkita shart o'rtasida namunalarni solishtirish mumkin bo'lgan Robust Rank Aggregation ('MAGeCK-RRA') yordamida original MAGECK algoritmi bilan solishtirganda, MAGECK-MLE murakkab eksperimental dizaynlarni modellashtirishga qodir. Bundan tashqari, MAGECK-MLE sgRNK nokaut samaradorligini aniq modellashtiradi, bu turli xil ketma-ketlik tarkibiga va xromatin tuzilmalariga qarab farq qilishi mumkin [11, 12]. MAGECK-MLE da sgRNKni o'qish soni i maqsadli gen g namunada j manfiy binom (NB) tasodifiy o'zgaruvchisi sifatida modellashtirilgan. NB taqsimotining o'rtacha qiymati (m ij) uchta omilga bog'liq: namunaning ketma-ketlik chuqurligi j (s j), sgRNKning nokaut samaradorligi i, va turli sharoitlarda (ya'ni, turli dori-darmonlarni davolash) genga ta'sirning chiziqli kombinatsiyasi g. Agar sgRNA bo'lsa i maqsadli genni yo'q qiladi g samarali, keyin m ij quyidagicha modellashtirilgan:

Ning ta'siri r turli shartlar ball sifatida ifodalanadib gr', differensial ifoda tahlilida "log qatlam o'zgarishi" atamasiga o'xshash gen tanlovlarining o'lchovi. Har bir namunadagi har bir shartning mavjudligi yoki yo'qligi ikkilik elementlariga kodlangan dizayn matritsasi d jr, va tajriba dizaynlaridan olinishi mumkin. 'bBallar har bir shartdagi tanlov darajasini aks ettiradi: b gr >0 (yoki <0) degani g holatda ijobiy (yoki salbiy) tanlangan r. m ij ga ham bog'liqdir b i0, odatda plazmidda yoki tajribaning 0 kunida o'lchanadigan dastlabki sgRNK ko'pligi.

ning qiymatlari b, sgRNKning samarali ekanligi haqidagi ma'lumot bilan birgalikda barcha sgRNK o'qish sonlarini kuzatishning qo'shma log-ehtimolini maksimal darajada oshirish orqali baholanishi mumkin. g barcha turli namunalarda va kutish-maksimizatsiya (EM) algoritmi yordamida optimallashtirilgan. EM algoritmida MAGECK-MLE iterativ ravishda har bir sgRNKning nokaut samaradorligini joriy baholash asosida aniqlaydi.b"Ballar (E bosqichi) va qayta hisoblash uchun nokaut samaradorligi to'g'risidagi yangilangan ma'lumotlardan foydalaniladi"bball (M bosqichi). Barcha namunalar bo'yicha har bir sgRNKni o'qishni hisoblash naqshlarini o'rganish orqali EM algoritmi samarasiz sgRNKlarning ta'sirini kamaytiradi. Usulning batafsil tavsifi Usullar bo'limida keltirilgan.

Biz MAGECK algoritmlarini to'rtta ommaviy ma'lumotlar to'plamida sinab ko'rdik. Birinchi ikkita ma'lumotlar to'plami ("ESC" va "leykemiya" ma'lumotlar to'plami) mos ravishda sichqonchaning embrion ildiz hujayralari (ESC) va ikkita inson leykemiya hujayra liniyasi (KBM7 va HL-60) bo'yicha salbiy tanlov tajribalariga mos keladi (3a va b-rasm). [1, 4]. Ikkala ma'lumot to'plamida ham hujayralar tabiiy o'sish holati bilan o'stirildi va CRISPR/Cas9 faollashtirilgandan so'ng hujayralarda salbiy tanlovlar paydo bo'ldi. Boshqa ikkita ma'lumotlar to'plami ("melanoma" nokauti va faollashuv ma'lumotlar to'plami) BRAF V600E mutatsiyasini o'z ichiga olgan A375 inson melanoma hujayralari liniyasidagi turli CRISPR ekranlaridir (4 va 5-rasm). Hujayralar BRAF inhibitori vemurafenib (PLX) yoki dimetil sulfoksid (DMSO) nazorati bilan ishlov berildi va GeCKO [2] yoki CRISPR/dCas9 Sinergetik Aktivatsiya Mediatori (SAM) kutubxonalari [5] bilan skrining qilindi. Ushbu ikkita ma'lumotlar to'plami original MAGECK-RRA algoritmi yordamida to'g'ridan-to'g'ri solishtirish qiyin bo'lgan bir nechta eksperimental sharoitlarni o'z ichiga oladi. Melanoma nokaut ma'lumotlar to'plamida hujayralar 7 yoki 14 kunlik tanlov ostida edi [2]. Melanoma faollashuvi ma'lumotlar bazasida lentiviral infektsiyasi bo'lgan hujayralarni tanlash uchun ikki xil dori (puromitsin va zeotsin) ishlatilgan, DMSO va PLX muolajalari 3 kunlik yoki 21 kunlik tanlov ostida profillangan [5].

Gen muhimligi ballari (b ball) MAGECK-MLE dan ikkita shart bo'yicha xabar berilgan. a, b the b Leykemiya ma'lumotlar bazasida ikkita leykemiya hujayralari chizig'ining ballari (a) va ESC ma'lumotlar to'plamidagi sichqonchaning ESC hujayralarining ikkita biologik nusxasi (b). Ichida (a), ba'zi taniqli haydovchi genlar va hujayra turiga xos genlar ham etiketlanadi. Ushbu genlar ikki xil leykemiya subtiplarida (HL60: o'tkir miyeloid leykemiya KBM7: surunkali miyeloid leykemiya), shu jumladan HL60 uchun CDK6 va TRIB1 va KBM7da RUNX1da alohida rol o'ynashi mumkin. AML o'sishida [17] CDK6 talab qilinadi va AML bilan og'rigan bemorlarda CML [18] bilan solishtirganda TRIB1 ning haddan tashqari ifodasi kuzatiladi. Boshqa tomondan, tez-tez RUNX1 funktsiyasini yo'qotish mutatsiyalari CML dan AMLga o'zgarishida kuzatiladi [19]. v Ikki shartli taqqoslash algoritmlari (masalan, RRA, ko'k to'rtburchaklar) bilan aniqlangan differentsial tanlangan genlarning tasviri. MAGECK-MLE hujayra turiga xos genlarni (qizil nuqta) boshqa genlardan ajrata oladi. Hujayra tipiga xos genlar bir holatda muhim ahamiyatga ega emas, lekin boshqa holatda kuchli muhimlikka ega bo'lgan genlardir va odatda biologik jihatdan qiziqroqdir.

The b melanoma nokaut ma'lumotlar to'plamida MAGECK-MLE ballari. a K-ning klasterlashgan ko'rinishini bildiradi b yuqori tanlangan genlardan barcha shartlar ballari (k = 4). Faqat eng yuqori yoki eng past 1% bo'lgan genlar b DMSO yoki PLX 14 kunlik davolash sharoitida ballar ko'rsatilgan. b Turli xil algoritmlardan foydalangan holda to'rtta shart bo'yicha ballarni taqsimlash. MAGECK-MLEdagi qizil to'rtburchak 4a-rasmdagi 4-klasterdagi genlarni yoki PLX 14 kunlik holatida kuchli ijobiy tanlangan genlarni bildiradi. Asl tadqiqotda ba'zi tasdiqlangan genlar qizil nuqta bilan belgilanadi, jumladan NF1, NF2, MED12 va CUL3.

The b melanomani faollashtirish ma'lumotlar to'plamida MAGECK-MLE ballari. a K-ning klasterlashgan ko'rinishini bildiradi b yuqori tanlangan genlardan barcha shartlar ballari (k = 5). Faqat eng yuqori yoki eng past 1% bo'lgan genlar b DMSO yoki PLX 21 kunlik davolash sharoitida ballar ko'rsatilgan. b o'rtacha b A ning 5-klasteridagi genlar ballari (zeotsin va puromitsin sharoitida ham izchil ijobiy tanlangan genlar), shuningdek b melanoma nokaut ma'lumotlar to'plamida ushbu genlarning ballari. O'xshash (a), k-vositasi klasterlash algoritmi tanlangan genlarga qo'llaniladi (k = 4)

Ikki shartli taqqoslashda MAGECK-MLE MAGECK-RRA, RSA va RIGER kabi mavjud usullar bilan o'xshash natijalar beradi. Barcha algoritmlar har xil hujayra tiplari uchun muhim bo'lgan genlarni aniqladi [16], shuningdek ikkita melanoma ma'lumotlar to'plamida PLX bilan davolangan sharoitda tanlangan ijobiy tanlangan genlar aniqlandi (3 -rasm, shuningdek, 1 -faylning A va B bo'limlariga qarang). Leykemiya ma'lumotlar to'plamida ikki shartli taqqoslash algoritmlari (masalan, MAGECK-RRA) HL60 va KBM7 ni to'g'ridan-to'g'ri taqqoslash orqali ikkita hujayra chizig'ida differentsial ravishda tanlangan genlarni aniqladi (3a-rasm) [10]. Ammo, bu genlarning hammasi ham bir xil biologik jihatdan qiziq emas, chunki MAGeCK-MLE ularni yana ikkita guruhga ajratdi: birida kam ta'sir ko'rsatadigan genlar (b ballari nolga yaqin), lekin boshqa hujayra chizig'ida kuchli selektsiya ta'siri (katta mutlaq b ballari) ), va ikkita hujayra chizig'ida zaif va qarama-qarshi ta'sirga ega bo'lgan genlar (3c-rasm). Birinchi guruh genlar ko'pincha biologik jihatdan qiziqroq, chunki ular hujayra turiga xos genlardir. Bunga ba'zi taniqli haydovchi genlari (masalan, KBM7 da BCR) hamda faqat bitta hujayra turida ishlashi mumkin bo'lgan genlar kiradi: HL60 da CDK6 va TRIB1 [17, 18] va KBM7 da RUNX1 [19].

MAGECK-MLE ning boshqa usullarga nisbatan afzalliklaridan biri shundaki, u bir ishda bir nechta shartlar va eksperimentlar bo'yicha gen muhimligini aniq taqqoslash imkonini beradi (4-rasm va 1-qo'shimcha faylning C bo'limi). Melanoma nokauti ma'lumotlar to'plamida, eng yaxshi tanlangan genlarning b balllarining k-ko'rsatkichlari klasteri bu genlar turli xil sharoitlarga ega ekanligini ko'rsatdi (4a-rasm). Ba'zi genlar barcha sharoitlarda universal ijobiy yoki salbiy tanlanadi (3-klaster), boshqalari esa turli sharoitlarda (1, 2 va 4-klasterlar) har xil ahamiyatga ega. 4-klasterdagi genlar ayniqsa qiziqarli, chunki ular 14 kunlik PLX bilan davolangan holatda kuchli ijobiy tanlovni ko'rsatadi. Darhaqiqat, nokauti PLX bilan davolash qilingan hujayralarda kuchli ijobiy tanlovga olib keladigan genlar 4-klasterda, jumladan NF1, NF2, MED12, CUL3 [2]. Bundan farqli o'laroq, boshqa algoritmlardan o'lchovlarning k-o'rtacha klasterlari 4-klasterdagi genlarning kuchli ta'sirini aniqlamadi (1-qo'shimcha faylning C bo'limi). Buning sababi shundaki, ularning bal taqsimotlari har xil sharoitlarda o'xshash (4b-rasm) va bitta shart (PLX 14 kunlik davolash) boshqa sharoitlarga qaraganda ancha kuchli ijobiy tanlovni keltirib chiqarmaydi [2]. Bu qisman MAGECK-RRA, RIGER va RSA ikkita shart o'rtasida sgRNAni solishtirish uchun darajaga asoslangan usuldan foydalanganligi sababli, miqdoriy ma'lumotni yo'qotishi mumkin.

MAGECK-MLE-ni bir nechta sharoitlarda qo'llashning yana bir misoli melanomani faollashtirish ma'lumotlar to'plamida ko'rsatilgan, bu erda hujayralar turli xil tanlash usullari (puromisin yoki zeotsin yordamida), dori-darmonlarni davolash (DMSO yoki PLX) va davomiyligi (3 kunlik yoki 21 kunlik davolash) o'tkazilgan. ) (5-rasm). Melanoma nokautining ma'lumotlar to'plamiga o'xshab, biz eng yaxshi tanlangan gen b ballarini k-o'rtacha klasterlashni amalga oshirdik. Ko'p ijobiy tanlangan genlar biologik jihatdan qiziq bo'lmasligi mumkin bo'lgan tanlov usuliga bog'liq. Misol uchun, 2 va 4-klasterlardagi genlar mos ravishda puromisin yoki zeotsin tanloviga xos bo'lgan ijobiy tanlangan genlarga mos keladi. Kichik genlar to'plami (5-klaster) zeotsin va puromisinda izchil tanlanadi, shu jumladan asl tadqiqotda tasdiqlangan genlar, masalan, EGFR, GPR35, LPAR1/5 [5]. Biz yana 5-klasterdagi genlarni ko'rib chiqdik (5b-rasm) va CRISPR faollashtirish tajribasida ijobiy tanlangan, ammo nokaut tajribasida kuchli salbiy tanlangan genlarga e'tibor qaratdik. Ushbu genlar EGFR va BRAF, melanoma rivojlanishini va PLX qarshiligini qo'zg'atuvchi ikkita ma'lum kinazlarni va RAS va JUN yo'lini faollashtiradigan protein kinaz CRKLni o'z ichiga oladi. CRKL kuchaytirilishi EGFR ning quyi oqim yo'llarini faollashtirish orqali EGFR inhibitörlerine qarshi dori qarshiligiga olib keladi [22], bu uning PLX dori qarshiligidagi potentsial rolini anglatadi.

VISPR yordamida QC o'lchovlari va genlarning muhimligini vizualizatsiya qilish

VISPR (crisPR ekranlarining VISualizatsiyasi) bu CRISPR QC ekranining interaktiv vizualizatsiyasi va taqqoslash natijalari uchun veb-ga asoslangan frontend. Interfaol kirish HTML5 asosidagi brauzer interfeysi bilan ta'minlanadi, vizualizatsiya esa Data-Driven Documents (D3) [24] ustidagi deklarativ vizualizatsiya grammatikasi Vega [23] yordamida amalga oshiriladi. VISPR CRISPR skriningini interaktiv tadqiq qilish uchun uch turdagi ko'rinishlarni taqdim etadi: sifat nazorati ko'rinishi, natijalar ko'rinishi va tajriba taqqoslash ko'rinishi. Sifat nazorati ko'rinishida avval tasvirlangan QC o'lchovlari ko'rsatilgan (2-rasm).

Natija ko'rinishida skrining natijalari interaktiv tarzda o'rganilishi mumkin. Unda har bir genning taqqoslash natijalari ko'rsatilgan jadval mavjud (6a-rasm). Jadvalni turli ustunlar bo'yicha saralash va gen nomlari yoki oddiy iboralar orqali filtrlash mumkin ("Qidiruv" dan). Bundan tashqari, tarqatish P qiymatlar kümülatif tarqatish funktsiyasi (CDF) (6b -rasm) va gistogramma (6c -rasm) sifatida ko'rsatiladi. Har bir gen uchun, barcha namunalarda normallashtirilgan sgRNK ko'rsatkichlari parallel koordinatali vizualizatsiya ko'rinishida ko'rsatilishi mumkin (6d -rasm). Agar mavjud bo'lsa, nokaut samaradorligi prognozlari [11] va har bir sgRNKning gen koordinatalari alohida o'qlar sifatida ko'rsatiladi. O'qlarni tartibini o'zgartirish yoki yoqish yoki o'chirish mumkin, va har bir o'qda diapazonlarni tanlash orqali sgRNKlarni ajratib ko'rsatish mumkin. Jadvalda tanlangan genlar CDFda ta'kidlangan bo'lib, ularning ichida paydo bo'lishini baholashga imkon beradi P barcha genlarning qiymat taqsimoti.

VISPR natijalari va taqqoslash ko'rinishlari. Natija ko'rinishida genlarni taqqoslash jadvali mavjud (a), taqsimlanishi P qiymatlari CDF (b) va gistogramma (v) va normallashtirilgan sgRNK tanlangan genlarning barcha namunalarida xromosoma joylashuvi va prognoz qilingan samaradorlik bilan (d). VISPR ning taqqoslash ko'rinishi (e) turli xil tanlanadigan sharoitlarda va eksperimentlarda muhim genlar o'rtasidagi o'zaro bog'lanishni Eyler diagrammasi sifatida ko'rsatadi. Ko'rsatilgan natijalar ESC (a, d) va melanoma ma'lumotlar to'plami (b, v, e)

VISPR tahlil natijalarini qo'shimcha o'rganishning turli usullarini taqdim etadi. Individual genlarni Ensembl [25] va IGV [26] da koʻrish mumkin. Tanlangan genlarni o'zaro ta'sir tarmog'i va funktsiyasi jihatidan GeneMANIA [27] orqali ko'rish mumkin. Funktsional tahlil GOrilla [15], onlayn Gen Ontologiyasi (GO) boyitish tahlili vositasi yordamida amalga oshirilishi mumkin. GOrilla genlarning reytingli ro'yxatini oladi (bu erda P MAGeCK tomonidan bildirilgan qiymatlar) boyitishni chegarasiz tahlil qilish uchun. Natijada paydo bo'lgan GO atamasi boyitishlari gen darajasida sifat nazorati uchun ishlatilishi mumkin.

VISPRning taqqoslash ko'rinishi Euler diagrammasi orqali umumiy va eksklyuziv muhim genlarni vizualizatsiya qilish orqali turli tajribalarni solishtirishi mumkin (6e-rasm). Eyler diagrammasining segmentlarini bosish tegishli tajribalarning natija ko'rinishini ochadi. Misol uchun, ikkita tajriba o'rtasidagi kesishuvni bosish har bir tajriba uchun "cheklangan" natijalar ko'rinishini ochadi, bu erda faqat umumiy muhim genlar ko'rsatiladi. Bu ko'rinishlar yuqorida tavsiflangan cheklanmagan natija ko'rinishlari bilan bir xil xususiyatlarni beradi. Biroq, bu holda, GOrilla bilan boyitish tahlili ko'rsatilgan genlar (ya'ni kesishgan genlar) oldingi planda va tajribaning boshqa genlari fon sifatida amalga oshiriladi.

VISPR-da ko'rsatilgan vizualizatsiyalarni nashrga tayyor SVG fayllari sifatida yuklab olish mumkin. Bundan tashqari, vizualizatsiyani Vega spetsifikatsiyasi sifatida saqlash uchun buyruq qatori interfeysi taqdim etiladi. Ushbu format foydalanuvchilarga VISPR chiqishini dasturiy tarzda o'zgartirish va uslublash imkonini beradi.

Snakemake bilan MAGECK-VISPR ish jarayonini amalga oshirish

Biz Snagemake [13] ish oqimini boshqarish tizimi bilan MAGeCK-VISPR ish oqimini amalga oshirdik, bu MAGeCK-VISPR funktsiyalarining bir qismini yoki barchasini avtomatik ravishda bajarishga imkon berdi: sifat nazorati, muhim genlar tahlili va vizualizatsiya. Snakemake kabi ish oqimini boshqarish tizimini tanlash bir qator afzalliklarga ega. Birinchidan, ish jarayoni bosqichlari avtomatik ravishda parallellashtirilishi va ish oqimini o'zgartirmasdan ish stantsiyalari, serverlar va hisoblash klasterlarida bajarilishi mumkin. Ikkinchidan, Snakemake barcha yaratilgan natija va oraliq fayllar uchun metamaʼlumotlarni (masalan, yaratilgan sana, kiritilgan va jurnal fayllari) kuzatib boradi. Shunday qilib, ishlatilgan ma'lumotlar, usullar va parametrlar har bir tahlil uchun (shuningdek, ma'lumotlarning kelib chiqishi deb ataladi) har tomonlama hujjatlashtiriladi, bu takrorlanadigan fanning muhim talabidir. MAGECK-VISPR ma'lum bir ish katalogida ish jarayonini ishga tushirish uchun buyruq qatori interfeysini taqdim etadi. Bu ish oqimi ta'rifini shunday deb o'rnatadi Ilon fayli, konfiguratsiya fayli va hujjatlar bilan birga. Konfiguratsiya fayli MAGECK-VISPR uchun xom ma'lumotlarning joylashuvi va qo'shimcha parametrlarni aniqlash uchun ishlatiladi. Konfiguratsiya qilinganidan so'ng, Snakefile Snakemake yordamida bajarilishi mumkin. Ilon fayli berilgan ish katalogiga o'rnatilganligi sababli uni foydalanuvchi osongina o'zgartirishi yoki kengaytirishi mumkin.

Biz ish jarayonining barcha komponentlarini Conda paketlari [28] sifatida taqdim etamiz, shunday qilib MAGECK-VISPR bitta buyruq bilan o'rnatilishi mumkin. Majburiy emas, Conda paket menejeri ish oqimi uchun alohida muhit yaratishi mumkin, masalan, dasturiy ta'minotning turli xil versiyalarini muzlatish yoki taqqoslash, yoki MAGeCK-VISPR ish oqimi nusxasini barcha ma'lumotlar va ishlatilgan dasturiy ta'minot bilan birga chop etish. Bu hosil qilingan natijalarning takrorlanishini yanada oshiradi.


Fakultet

Ilmiy qiziqishlar: Mening tadqiqotim kompyuter modellaridan foydalanish orqali biologik tizimlarni yaxshiroq tushunishga qaratilgan hisoblash tizimlari biologiyasi sohasida. Mening faol tadqiqot yo'nalishlarim orasida mashhur Gepasi simulyatori muallifi va COPASI simulyatori rahbari (U. Kummer bilan) sifatida modellashtirish va simulyatsiya dasturlarini ishlab chiqish kiradi va SBML, tizim biologiyasini belgilash tili va va model annotatsiyasi uchun MIRIAM taklifi. Mening guruhim shuningdek, biokimyoviy modellarni quradi, hozirda bu temir metabolizmi, eukaryotik translatsiya va mikrobial markaziy metabolizm modellarini o'z ichiga oladi. Ushbu ish orqali men biokimyoviy kinetik modellashtirishda raqamli global optimallashtirishni qo'llashda kashshof bo'ldim va tizim biologiyasida, xususan, ma'lumotlardan teskari muhandislik modellari uchun rasmiy tizimlarni identifikatsiyalash usullaridan foydalanishga qiziqaman. Mening tadqiqotim keng fanlararo yondashuvni talab qiladi va men o'z tadqiqot guruhida yoki hamkor sifatida fanning ko'pgina sohalaridagi odamlar bilan ishlayman.

Laboratoriya a'zolari:

--Dr. Sherli Koshy-Chenthittayil (postdoktorant) aralash turdagi biofilmlarning modellarini ishlab chiqmoqda.

--Dr. Hasan Baig (doktorlikdan keyingi tadqiqotchi) COPASI dasturiy ta'minotini ishlab chiqish ustida ishlamoqda.

--Joe Masison (MD/PhD talabasi) temir almashinuvining ko'p miqyosli modellarini ishlab chiqmoqda.

--Aidan Riley (bakalavr) COPASI dasturiy ta'minotini ishlab chiqishga hissa qo'shmoqda.

Molekulyar, mikrobial va strukturaviy biologiya professori

Hujayra biologiyasi kafedrasi assistenti

Genetika va rivojlanish biologiyasi kafedrasi dotsenti
Blinov laboratoriyasi

Molekulyar, mikrobial va strukturaviy biologiya boʻyicha faxriy professor

Biotibbiyot muhandisligi kafedrasi assistenti
Tadqiqotchi, Yel saraton tizimlari biologiya markazi
Yel universitetining tashrif fakulteti

Ilmiy qiziqishlar: Laboratoriyamizning maqsadi hujayraning qo'shnilari va mikromuhit bilan o'zaro ta'sirini tushunishdir. O'sma mikromuhitiga alohida e'tibor qaratgan holda, biz til hujayralari bir-biri bilan suhbatlashish uchun foydalanadigan, til grammatikasi va mazmuni atrof-muhitga qanday moslashishi va hujayralar bunga javoban qanday harakat qilishini aniqlashni maqsad qilgan. Biz biologiya va tibbiyotning asosiy savollariga javob berish uchun jonli mikroskopiya, bioinformatika, to'qima muhandisligi, nanofabrikatsiya, hujayra namunasi, evolyutsion biologiya va genetika kabi bir qator texnikadan foydalanamiz. Bizning hozirgi fiziologik yo'nalishimiz saraton populyatsiyalarida yangi fenotiplarning qo'shni o'zaro ta'siri orqali metastatik tolerantlikning evolyutsion asosi va kasalliklarni modellashtirish uchun yurakning etukligi qanday paydo bo'lishini tushunishni o'z ichiga oladi.

Laboratoriya a'zolari:

--Yosir Suhayl saraton metastazining atrofdagi stromal muhitga bog'liqligi bilan shug'ullanuvchi postdoktorlik tadqiqotchisi.

--Visar Ajeti-doktorlikdan keyingi ilmiy xodim.

Hujayra biologiyasi kafedrasi professori

Kompyuter fanlari va muhandisligi professori

Vasiylar kengashi Hurmatli professor

Ilmiy qiziqishlar: Mening laboratoriyam hujayra funktsiyasining asosiy mexanizmlarini tushunishga yordam beradigan yangi eksperimental va hisoblash texnologiyalarini ishlab chiqishga qaratilgan. Bizda membrana potentsialining lyuminestsent problarini ishlab chiqish va tavsiflashga qaratilgan uzoq vaqtdan beri harakatimiz bor. Yaxshiroq sezgir kuchlanish ko'rsatkichlarini ishlab chiqish uchun organik kimyoviy dizayn va sintezdan foydalangan holda bu harakat davom etmoqda. Bu ish hujayra membranasi bo'ylab va sitoplazma ichida signalizatsiya yo'llarini tashkil etishga doimiy qiziqish uyg'otdi. Bizni hujayralardagi molekulalarning murakkab fazoviy tashkil etilishi hujayra funktsiyasini boshqarish uchun qanday qo'llanilishi haqidagi juda umumiy savol qiziqtiradi. Bu, ayniqsa, neyron hujayralari uchun to'g'ri keladi va biz miya to'qimalarida dendritik tikanlar hujayra biologiyasini faol ravishda o'rganmoqdamiz. Bu bizni "Virtual hujayra" deb nomlangan juda umumiy hisoblash modellashtirish dasturiy platformasini ishlab chiqishga undadi, unda biz hujayra biologik mexanizmlarini o'rganish uchun kompyuter simulyatsiyasidan foydalanish uchun asos yaratdik. Modellar tabiiy ravishda biokimyoviy va elektrofiziologik ma'lumotlar bilan birlashtirilgan hujayrali va hujayra osti tuzilmalarining eksperimental tasvirlaridan qurilgan. Biz "Virtual hujayra" tizimidan neyron va neyron bo'lmagan hujayralardagi elektr va signalizatsiya faolligini tushuntirish uchun foydalanganmiz. Yaqinda biz hujayradan tashqari miqyosda molekulyar o'zaro ta'sirlarni modellashtirishi mumkin bo'lgan SpringSaLaD nomli yangi dasturiy ta'minotni ishlab chiqdik. Bu bizga individual molekulalarning shakli ularning o'zaro ta'sirini qanday boshqarishini o'rganishga imkon beradi va natijada hujayra darajasidagi javoblarga olib keladi.

Laboratoriya a'zolari:

Professor va genetika va rivojlanish biologiyasi kafedrasi muovini

Tadqiqot Qiziqishlar: Bizning guruhimiz signallarni uzatish mexanizmlari bilan qiziqadi. Hujayraning atrofdagi muhitdan signallarni qabul qilish va bu signallarga to'g'ri javob berish qobiliyati tom ma'noda hayot va o'lim masalasidir. Hujayra ko'payadimi, farqlanadimi yoki o'ladimi, qaerga yopishishi yoki ko'chishi, uning xatti-harakatlarining deyarli barcha jihatlari signallarni to'g'ri talqin qilish qobiliyatiga bog'liq. Signal nafaqat normal rivojlanish va organizmning kundalik faoliyati uchun juda muhim, balki tartibga solinmagan signalizatsiya saraton va otoimmün kasalliklar kabi ko'plab inson kasalliklarining asosini tashkil qiladi. Endi ma'lum bo'lishicha, signalizatsiya mexanizmining markaziy elementlaridan biri oqsil-oqsil komplekslarining yuqori darajada tartibga solinadigan va o'ziga xos shakllanishi hisoblanadi. Signalning oqsillarning bir-biriga bog'lanishiga tayanishi ajoyib imkoniyatlarni taqdim etadi: bog'lanish signal yo'llarining muhim tarkibiy qismlarini aniqlash vositasi sifatida ishlatilishi mumkin, shuningdek, laboratoriya yoki klinikada ushbu yo'llarni inhibe qilish strategiyalari uchun asos bo'lib xizmat qiladi. Biz biokimyoviy va hujayra biologik usullarining kombinatsiyasidan foydalanamiz, masalan, hujayra proliferatsiyasi va sitoskeletning tashkil etilishini boshqaruvchi signal yo'llarini tushunamiz. Biz, shuningdek, funktsional jihatdan muhim oqsil o'zaro ta'sirini aniqlash va global miqyosdagi o'zaro ta'sirlarni tavsiflash uchun yangi proteomik yondashuvlarni faol ravishda davom ettirmoqdamiz.

Laboratoriya a'zolari:

--Kazuya Machida laboratoriyada yashovchi dotsent bo'lib, hujayralardagi SH2 domenini bog'laydigan joylardan saraton biomarkeri sifatida foydalanish uchun asboblar ishlab chiqaradi.

--Grace Curley-Jons - UConn bakalavriati bo'lib, doktor Machida bilan SH2 domenini xaritalash bo'yicha ishlaydi.

Hujayra biologiyasi professori

Hujayra biologiyasi professori

Tadqiqot qiziqishlari: Mening laboratoriyamdagi tadqiqotlar hujayra ichidagi transportning molekulyar mexanizmlari va mikrotubulalar sitoskeletini tashkil etishga qaratilgan. Melanoforlar, pastki umurtqali hayvonlarning pigment hujayralari, katta hujayralar bo'lib, ular bir vaqtning o'zida membrana bilan chegaralangan minglab pigment granulalarini qattiq agregat hosil qilish yoki sitoplazma bo'ylab bir tekis tarqalish uchun tezda hujayra markaziga olib boradi. Agregatsiya jarayonida granulalar sitoplazmatik dinein yordamida mikronaychalar bo'ylab harakatlanadi, ammo dispersiya Kinesin II tomonidan mikrotubulaga bog'liq bo'lgan dastlabki transportni va keyinchalik tasodifiy joylashtirilgan aktin filamentlari bo'ylab sekin diffuziyaga o'xshash harakatni o'z ichiga oladi. Tashish Protein Kinaz A (PKA) signal kaskadi tomonidan tartibga solinadi va shu bilan hujayra ichidagi transportda sitoskeleton rolini, ikkita asosiy transport tizimi o'rtasida almashinish mexanizmlarini va signal uzatish orqali motor molekulalarining faolligini o'rganish uchun noyob tizimni ta'minlaydi. mexanizmlar. Hujayra ichidagi transportni tartibga solish asosida yotgan molekulyar mexanizmlarni tushunish uchun biz molekulyar biologiya va biokimyodan, jonli hujayra floresan mikroskopiyasidan, fotooqartirishdan, fotoaktivatsiyadan va motorga xos zondlarning mikroin'ektsiyasidan foydalanamiz.

Hujayra biologiyasi kafedrasi dotsenti

Hujayra biologiyasi kafedrasi dotsenti

Ilmiy qiziqishlar: Bizning tadqiqotimiz hisoblash tizimlari tibbiyoti va tizim biologiyasi, matematik biologiya va bioinformatika chorrahasida joylashgan. Biz biologik tizimlarni modellashtirish, simulyatsiya qilish va boshqarish uchun matematik algoritmlarni loyihalash, dasturiy ta'minotni ishlab chiqish va qo'llash ustida ishlaymiz. Molekulyar biologiyada bizni qiziqtirgan tizimlar genlarni tartibga soluvchi tarmoqlar va hujayra ichidagi signalizatsiya tarmoqlarini o'z ichiga oladi, bu erda biz hujayralarning murakkab tartibga solish dasturlarini tushunish va nazorat qilishni maqsad qilganmiz. We are focused on Cancer research (cancer reversion mechanisms and reversion of chemotherapy resistance) in breast cancer and leukemia.

Lab members:

--Lauren Marazzi is a MD/PhD student working on quantitative analysis of chemotherapy resistance reversion and cancer reversion in triple negative breast cancer using mathematical modeling and control theory.

Associate Professor of Genetics and Developmental Biology

Research Interests: Research in our laboratory focuses on developing quantitative imaging tools that reveal dynamics of cellular signaling at high spatial and temporal resolution (biosensors), or that enable optical control of signaling proteins at with temporal and spatial precision (optogenetics). These tools are used with live cell microscopy to understand signaling networks underlying cellular function including cell polarity cell motility, axon guidance and development of dendritic spines in neurons. In building biosensors we use structural design strategies with fluorescent proteins to optimize FRET-based activity reporters for signaling proteins. Multiple live cell imaging modalities are used in the lab including TIRF, confocal, intensity-based ratiometric imaging, and fluorescence lifetime microscopy (FLIM). Optogenetics has generated much interest as it allows precise control of signaling in living systems using genetic engineering of natural photosensory proteins. We are exploring the use of the flavin-binding LOV (light-oxygen-voltage) domain from the plant photoreceptor phototropin. We used this technology to produce a genetically-encoded photoactivatable analog of Rac (PA-Rac) that enables precise modulation of Rac activity at regions that are submicrons in size and capable of controlling activation with microseconds precision in living cells. Current projects in the lab focus on extending such technology to other signaling proteins.

Lab members:

--Yuezhe Li is a PhD student working on the role of primary cilia on insulin signaling of pancreatic beta cells.

Assistant Professor of Cell Biology

Research Interests: My lab is focused on developing membrane potential probes to image neuronal and cardiac activities. We use synthetic organic chemistry, genetic techniques, optical spectroscopy and microscopy to decode cells' secrets.

Lab members:

Associate Professor of Genetics and Developmental Biology

Lab Website: Yu Lab

Research Interests: The Yu lab is interested in developing new microscopy imaging techniques. Currently our works are focused on two related areas: (1) Live cell single-molecule dynamics. We are working on techniques that allows better visualization of individual biomolecules in a living specimen using modern fluorescence microscopy techniques. We are particularly interested in understanding how protein molecules’ behavior changes during signal transduction, and how individual proteins are made during gene expression. See the following publications from more details: Das et al. PNAS 112(3), E267 Oh et al., PNAS 109(35):14024. (2) Super-resolution microscopy. For a long time fluorescence microscopy had suffered from the curse of the so-called “diffraction-limited resolution”, which sets a fundamental limit on what is the smallest object the microscope can see. In the last decade, however, many new forms of microscopy design have emerged with the goal of breaking this resolution limit. Collectively, they are often called “super-resolution optical microscopy”. The lab is actively working on this new field trying to further improve the imaging quality of these new technologies. For more detail, check out these recent papers: Yu J et al. PNAS 201912634 Elmokadem A et al. Biophysical J 109(9):1772

Lab Members:

--Nizam ud Din is a postdoc fellow working on developing SH2 imaging techniques to understand the spatial distribution of phosphoproteome in cells during signal transduction.

--Meagan Cauble is a postdoc fellow studying the mechanisms of the growth factor pathways in vertebral disk in response to tissue damage.


Abstrakt

To discern how mechanical forces coordinate biological outcomes, methods that map cell-generated forces in a spatiotemporal manner, and at cellular length scales, are critical. In their native environment, whether it be within compact multicellular three-dimensional structures or sparsely populated fibrillar networks of the extracellular matrix, cells are constantly exposed to a slew of physical forces acting on them from all directions. At the same time, cells exert highly localized forces of their own on their surroundings and on neighboring cells. Together, the generation and transmission of these forces can control diverse cellular activities and behavior as well as influence cell fate decisions. To thoroughly understand these processes, we must first be able to characterize and measure such forces. However, our experimental needs and technical capabilities are in discord—while it is apparent that we should study cell-generated forces within more biologically relevant 3D environments, this goal remains challenging because of caveats associated with complex “sensing–transduction–readout” modalities. In this Review, we will discuss the latest techniques for measuring cell-generated forces. We will highlight recent advances in traction force microscopy and examine new alternative approaches for quantifying cell-generated forces, both of individual cells and within 3D tissues. Finally, we will explore the future direction of novel cellular force-sensing tools in the context of mechanobiology and next-generation biomaterials design.


12918_2011_787_MOESM1_ESM.ZIP

Additional file 1: ML-Rules demo program The ZIP file comprises a prototype tool of ML-Rules including a model editor, the simulator, and a rudimentary line chart visualization of simulation trajectories. Also a user manual and several example models are part of the tool package. To start the demo tool, please unzip the file and execute the run.jar file. Java Runtime Environment (Version 6 or higher) is required for execution. (ZIP 6 MB)

12918_2011_787_MOESM2_ESM.PDF

Additional file 2: Example models The PDF file contains descriptions of the entire example models including initial solutions and parameter values that have been used for the simulation studies. (PDF 560 KB)


Spatial modelling of brief and long interactions between T cells and dendritic cells

In the early phases of an immune response, T cells of appropriate antigen specificity become activated by antigen-presenting cells in secondary lymphoid organs. Two-photon microscopy imaging experiments have shown that this stimulation occurs in distinct stages during which T cells exhibit different motilities and interactions with dendritic cells (DCs). In this paper, we utilize the Cellular Potts Model, a model formalism that takes cell shapes and cellular interactions explicitly into account, to simulate the dynamics of, and interactions between, T cells and DCs in the lymph node paracortex. Our three-dimensional simulations suggest that the initial decrease in T-cell motility after antigen appearance is due to ‘stop signals’ transmitted by activated DCs to T cells. The long-lived interactions that occur at a later stage can only be explained by the presence of both stop signals and a high adhesion between specific T cells and antigen-bearing DCs. Furthermore, our results indicate that long-lasting contacts with T cells are promoted when DCs retract dendrites that detect a specific contact at lower velocities than other dendrites. Finally, by performing long simulations (after prior fitting to short time scale data) we are able to provide an estimate of the average contact duration between T cells and DCs.

Fayl nomi Tavsif
imcb7100054-sup-0001.mpgapplication/mpg, 2 MB Supplementary Material
imcb7100054-sup-0002.mpgapplication/mpg, 2 MB Supplementary Material
imcb7100054-sup-0003.mpgapplication/mpg, 2 MB Supplementary Material
imcb7100054-sup-0004.mpgapplication/mpg, 2 MB Supplementary Material
imcb7100054-sup-0005.mpgapplication/mpg, 2 MB Supplementary Material
imcb7100054-sup-0006.docapplication/doc, 14 KB Supplementary Material

E'tibor bering: nashriyot mualliflar tomonidan qo'llab -quvvatlanadigan ma'lumotlarning mazmuni yoki funktsionalligi uchun javobgar emas. Har qanday so'rovlar (etishmayotgan kontentdan tashqari) maqola uchun tegishli muallifga yo'naltirilishi kerak.


Challenges and future directions

The accuracy of agent-based simulations relies on both the agents and virtual environment capturing key features and processes necessary for the emergence of a required collective behaviour. These are not always well understood and so close integration with biologists developing cellular models is essential to ensure that key agent behaviours and environmental factors are present. The Synthetic Biology Open Language (SBOL) [68] and SBML [69] are standards to aid in the exchange of genetic design information and unambiguous definition of biochemical models. Having agent-based tools exploit these formats directly would enable existing curated intracellular models to drive agent behaviours. This would help to validate their function when exposed to realistic extracellular factors, and provide clearer links between model parameters of relevance to the cell biology and desired population-level features. Furthermore, the integration of tools designed to efficiently model the reaction networks inside cells (e.g. Smoldyn [70] or NFsim [71]), and the application of whole-cell models [72] to provide detailed behavioural responses would enable accurate simulations. At present, most tools do not provide these features due to the extensive computational demands of simulating large and complex multi-scale models. However, as the availability of cheap high-performance computing grows, and agent-based tools are updated to better exploit these resources, large multi-scale modelling will become viable.

Many real-world applications of synthetic biology require cells to robustly function within complex environments. Faithfully representing key aspects of these environments is essential to ensure that simulations produce accurate results. The use of microfluidics to study single-cell dynamics has seen significant growth in synthetic biology [73]. Such devices impose intricate boundaries on cells that both physically restricts their movement and controls the flow of nutrients sustaining them. Although the role of fluid flows on natural biofilms has been investigated [74], there is a lack of agent-based modelling tools that incorporate the full range of physical processes that might be experienced by a cell, hampering the ability for them to fully describe many systems of this type.

A significant challenge when capturing the complexity of cellular populations is the typical number of individuals involved. Colonies of bacteria will far exceed 100 million cells. At this size, if only the position of each cell is maintained, over 1 GB of raw data would need to be updated for each time point of a simulation. The execution of models at these scales requires the adoption of efficient parallelizable algorithms and high-performance computing architectures. These allow for a simulation to be broken down into many smaller parts and large numbers of processing units used to solve each concurrently. A shift to highly parallel computing architectures has already taken place in molecular dynamics simulations, leading to huge leaps in the speed and scale of problems that can be solved [75]. Some attempts have also been made to use this approach for synthetic biology applications, e.g. CellModeller [63] exploits graphics processing units (GPUs) to accelerate simulations, but these optimizations often come at the cost of limiting the range of possible agent behaviours and the complexity of the virtual environment. While several of the general-purpose modelling frameworks (e.g. FLAME and Repast) do support these types of large-scale simulation, they also lack the biologically relevant built-in features (e.g. cell growth and simulation of genetic networks) that are critical for the efficient development of synthetic biology-related simulations. To further mitigate some of these computational difficulties, attempts have also been made to employ alternative forms of modelling. Hybrid approaches in which an agent-based model is combined with continuous models has been shown to significantly reduce the computational demands of some forms of simulation [76], and dynamic network-based models can be used to simplify the virtual environment, while still ensuring that interactions between cells are fully captured [77–80].

The large number of agent-based modelling tools raises the question: why do so many exist? This is partially due to historic reasons. As various sub-fields of biology have applied agent-based models, they each have developed tools containing the specific features they require. Although this makes it easier for them to tailor models to their specific needs, it also leads to numerous tools all focused on slightly different problems. It is conceivable that a single tool could eventually encapsulate the functionality of all of these. Some efforts in this direction have already begun with Chaste and BSim being built around a ‘plug-n-play’ architecture where simulations are built from a set of available modules. Because users can also define their own modules from scratch, the functionality of the tool can be easily extended in new ways. Intuitively, it would seem that this type of approach will eventually become the standard. However, this flexibility makes it impossible to highly optimize the interactions between modules. This results in less efficient simulations. Therefore there is always likely to be a range of modelling tools available, especially for specific areas that require the highest performance simulations.

In summary, our knowledge of the inner workings of cells has grown significantly over recent years. This has supported the development of genetically engineered cells able to sustainably produce useful chemicals [4] and implement novel behaviours [1–3,27,28,36,81–83]. Nevertheless, synthetic biology has struggled to effectively scale systems beyond individual cells to the rational engineering of multicellular collective functions. Agent-based modelling offers a way to explore the links between single-cell behaviours and population-level phenomena [9]. This will help to support the next wave of synthetic biology applications that exploit large populations of cells to implement robust functionalities at scale.


MAQDAT

We are grateful to two anonymous reviewers for their suggestions, as well as to Y. Caraglio (CIRAD) for his comments on some aspects of plant architecture. We thank LIAMA, the French Embassy in China, the National Natural Science Foundation of China (NSFC, # 60073007, 60473110), the Chinese Academy of Forestry, as well as the French research institutes CIRAD, INRA and INRIA for their support and funding of the organization of PMA06. We are very grateful to the editors at Annals of Botany for their kind co-operation in editing this Special Issue on plant growth modelling and applications.