Research Pulse — 2026-08-30 | Brown Biotech

# Research Pulse — 2026-08-30
_Generated 2026-08-30 15:01 KST · AI × Bioresearch Daily synthesis_

Brown Biotech daily research pulse — multi-front synthesis across CRISPR /
gene editing, AI drug discovery, protein AI, single-cell foundation models,
FDA/regulatory, and spatial transcriptomics. Sources: company filings,
peer-reviewed journals, FDA guidance, conference disclosures.

---

### Post 1: 오늘의 전체 흐름

🧬 **AI × Bioresearch Daily — 8월 30일 (일요일) · Mapping(매핑 Day)**

8/30(매핑)→8/29(Assessing)→8/27(Converging)→8/26(Translating)→8/25(Optimizing)→8/24(Transitioning)→8/23(Validating)→8/22(Assessing)→8/21(Scanning)→8/20(Expanding)→...→7/14(가속). Assessing(8/29)된 기술들이 **대규모 시공간적 매핑(mapping)으로 데이터베이스화하고 통합 가능성을 재확인하는 날** — SCLC 105개 patients · 600K cells 이상에서 lymph node metastasis의 공간적 세포 생태계 매핑(Zhang Cell Reports Medicine 2026), gastric cancer 4개 공간 영역에서 lymphocyte-aggregated region(LAR) 발현과 tertiary lymphoid structures(TLS) 확인(Gao Nature Communications 2026), FDA gene therapy 38개 CGT approvals 1998-2025 체계적 분석으로 규제 evolution과 future directions 도출(Oo Journal Clinical Pharmacology 2026), VCBench가 7개 dimension에서 single-cell foundation models의_virtual cell`_capability 평가(Weidener bioRxiv 2026), Feldman Nature Chemical Biology 2026)이 AlphaFold의 adversarial mutations에 대한 invariance를 정량 확인 — computational biophysics 한계 확인, pharma 2026년 AI platforms 대규모 딜 급증 확인(Lin GEN Biotechnology 2026)이 동시에 진행된다.

Mapping은 Assessing의 다음 단계다. 오늘 6개 전선은 **대규모 매핑 데이터로 기술의 통합 가능성과 한계를 동시에 재확인하는 날**이다.

*출처: [Zhang SCLC Spatial CosMx Cell Reports Medicine 2026](https://consensus.app/papers/details/dc8b02ddbfcb574fb92b18cf8aed1117/), [Gao Gastric Cancer LAR Nature Communications 2026](https://consensus.app/papers/details/46f3d71acd2e5b6e8816d6886aa485d9/), [Oo FDA Gene Therapy Approvals Journal Clinical Pharmacology 2026](https://consensus.app/papers/details/181658e93c825ea581a0664845a98d12/), [Weidener VCBench bioRxiv 2026](https://consensus.app/papers/details/48f053efc3485e6589bcd85cc6009842/), [Feldman AlphaFold Adversarial Computational Structural Biotechnology 2026](https://consensus.app/papers/details/68f35194db525fb68a7bc8c10546001b/), [Lin Pharma AI 2026 Deals GEN Biotechnology 2026](https://consensus.app/papers/details/bc6e9ea3ec2e5025af707131d2fa0ca7/)*

*#MappingDay #Zhang #SCLC #CosMx #600Kcells #LymphNodeMetastasis #Gao #GastricCancer #LAR #TLS #Oo #FDAGeneTherapy #1998-2025 #Weidener #VCBench #VirtualCell #Feldman #AlphaFold #Adversarial #Invariance #Lin #PharmaAI #2026Deals #Mapping #Assessing #SpatialTranscriptomics #PanCancer #CellularEcosystem #TumorMicroenvironment #NicheMapping #RegulatoryEvolution #CGT #FoundationModel #Benchmark #ProteinStructure #Biophysics*

---

### Post 2: CRISPR & Gene Editing — Nanomaterial-Mediated CRISPR/Cas Delivery의 최근 동향

🧪 **[CRISPR/Gene Editing] Nanomaterial-mediated CRISPR/Cas delivery가 lipid nanoparticles에서 vesicle-derived systems까지 확장 확인(Wang Frontiers Bioengineering 2026) + therapeutic genome editing의 delivery challenges/innovations로 ionizable lipid nanoparticles 임상 시험에서 93% protein knockdown 확인(Yang MedComm 2026) + CRISPR-Cas9 delivery strategies의 therapeutic applications 확장 확인(Makhijani Discover Nano 2026): CRISPR delivery의 nanomaterial 기반 확장(Expanding) 동시 확인**

CRISPR 유전자 편집 delivery의**nanomaterial 기반 확장**이 출처에서 동시에 확인된다[1][2][3].

**Nanomaterial-Mediated CRISPR/Cas Delivery: LNPs에서 Vesicle-Derived Systems까지[1]**
Bingning Wang et al.(Frontiers in Bioengineering and Biotechnology 2026)의 review는 CRISPR/Cas delivery를 위한 nanomaterials의 현재 상황을 종합한다. ** lipid-based·polymeric·inorganic·vesicle-derived systems 포함**: delivery efficiency·cell targeting·endosomal escape·intracellular movement 향상. **Ionizable lipid nanoparticles**: nucleic acids와 CRISPR systems 전달에 most advanced performance. **새로운 모달리티**: polymer-based·exosome-inspired carriers가 repeated·targeted applications 위해 빠르게 진행 중. **Hybrid·responsive systems**: better spatial·temporal control of editing 가능. **핵심**: nanomaterial-mediated CRISPR delivery가 lipid nanoparticles에서 vesicle-derived systems까지 확장 — CRISPR delivery의 nanomaterial 기반 확장.

**Therapeutic Genome Editing Delivery: Challenges · Innovations · 93% Protein Knockdown[2]**
Meijia Yang et al.(MedComm 2026)의 review는 therapeutic genome editing delivery의 현주소를 분석한다. **Viral vectors의 한계**: AAV의 restricted cargo capacity·immunogenicity·complex manufacturing. **Ionizable lipid nanoparticles의 성과**: hepatic targets에서 임상 시험 single dose 후 93% protein knockdown 달성. **새로운 모달리티**: virus-mimicking nanosystems·cell-derived extracellular vesicles·cell-penetrating peptides·intelligent-responsive multifunctional scaffolds. **ML 기반 vector optimization**: high-throughput barcoded screening과 machine learning이 vector optimization 가속화. **핵심**: ionizable LNPs로 93% protein knockdown 확인 + 새로운 delivery modalities 확장 — CRISPR delivery의 혁신적 성과.

**CRISPR-Cas9 Delivery Strategies: Therapeutic Applications 확장[3]**
Shivani Makhijani et al.(Discover Nano 2026)의 review는 CRISPR-Cas9 delivery strategies의 최신 동향을 분석한다. **Delivery approaches**: viral vectors·lipid nanoparticles·other nonviral vectors 개선으로 target specificity 향상·off-target effects 최소화. **Therapeutic applications 확장**: genetic disorders correction·immune cells engineering for cancer therapy·viral infections combating. **새로운 CRISPR variants**: base editing·prime editing·epigenome editing으로 accuracy 향상·intervention 범위 확대. **남은 도전**: unintended edits·immune responses·specific tissues/organs로의 delivery. 핵심: CRISPR-Cas9 delivery strategies로 therapeutic applications 확장 확인 — delivery 기술의 실용화 진행.

**8월 29일 Assessing Day vs 오늘의 차별점**: 8/29이 **Witten Nature Biotechnology**으로 **FO-32 · FO-35 deep learning-guided ferret lung nebulized delivery**를, **Chan Nature Nanotechnology**으로 **COMET transformer non-canonical LNP formulations**를 제시했다면, 오늘은 **Wang Frontiers Bioengineering**으로 **nanomaterial-mediated delivery의 전체 범위(LNPs + vesicle-derived systems) 종합**을(nanomaterial 확장), **Yang MedComm**으로 **ionizable LNPs로 93% protein knockdown 임상 성과**를(정량적 성과), **Makhijani Discover Nano**으로 **CRISPR-Cas9 delivery strategies의 therapeutic applications 확장**을(적용 영역)을 동시에 제시한다. Assessing(8/29) → nanomaterial 기반 delivery의 통합적 확장 + 정량적 성과 + 적용 영역 Mapping(8/30).

실무 함의: 유전자 치료 전달팀에서 Yang et al.의 93% knockdown 성과와 Wang et al.의 nanomaterial delivery 범위를 분석하여 delivery strategy 수립 시 hybrid approaches를 고려해야 한다.

*출처: [Wang Nanomaterial CRISPR Delivery Frontiers Bioengineering 2026](https://consensus.app/papers/details/f43692af0ccb54b684ef6714b3de8d82/), [Yang Genome Editing Delivery MedComm 2026](https://consensus.app/papers/details/87a151ebbb6759e48db9d0a028ff4597/), [Makhijani CRISPR Cas9 Delivery Discover Nano 2026](https://consensus.app/papers/details/7acc7eabb0f05c51bccf0d568ec55751/)*

*#CRISPR #GeneEditing #Wang #Nanomaterial #FrontiersBioengineering #VesicleDerived #Yang #MedComm #93PercentKnockdown #IonizableLNP #Makhijani #DiscoverNano #CRISPRCas9 #DeliveryPlatform #LipidNanoparticle #ExosomeInspired #VirusMimicking #MappingDay #NonViral #ViralVectors #AAV #EndosomalEscape #MachineLearning #VectorOptimization #TherapeuticApplications #BaseEditing #PrimeEditing #EpigenomeEditing #TargetedDelivery #GeneTherapy #ClinicalTranslation #CMC #Biocompatibility #Manufacturing #Regulatory*

---

### Post 3: AI 신약 — Pharma 2026 AI Deals 급증과 AI Drug Discovery의 현주소

💊 **[AI 신약] Pharma 2026년 AI platforms 대규모 딜 급증 확인(Lin GEN Biotechnology 2026) + AI drug discovery pipeline 전반 적용 종합(Ali Drug Development Research 2026) + AI drug discovery의 현재 한계와 기회 종합(Seyhan Critical Reviews Oncology/Hematology 2026): pharma AI 딜 급증과 AI drug discovery의 현실적 평가 Mapping 동시 확인**

Pharma AI 딜**급증과 AI drug discovery 현실 평가**가 출처에서 동시에 확인된다[4][5][6].

**Pharma 2026 AI Deals 급증: Flurry of 2026 Deals[4]**
Fay Lin(GEN Biotechnology 2026)의 분석: **pharma companies가 2026년 AI platforms에 대규모 투자**. Big pharma의 AI drug discovery partnerships 급증 확인. **AI platforms 거래 규모**: 2025년 대비 크게 증가. **주요 플레이어들**: multiple major pharmaceutical companies involved. **AI drug discovery 속도**: target identification·hit finding·lead optimization·drug repurposing 가속화. **향후 전망**: AI가 pharmaceutical innovation의 핵심 도구로 자리매김. 핵심: Pharma 2026 AI platforms 대규모 딜 급증 확인 — AI drug discovery의 산업적 중요성 재확인.

**AI Drug Discovery Pipeline 전반 적용: Transformative Innovation[5]**
M. Ali et al.(Drug Development Research 2026)의 review는 AI drug discovery의 pipeline 전반 적용을 분석한다. **전 단계 적용**: target identification·hit finding·lead optimization·drug repurposing에 AI 적용. **AI-driven in silico platforms**: toxicity·pharmacokinetics·developability 예측으로 late-stage attrition 감소. **Collaborative models**: AI developers와 pharmaceutical industries 간 협업 모델. **Challenges**: algorithmic transparency·data quality·interoperability·regulatory acceptance. **미래 방향**: personalized medicine·pharmaceutical formulation development. 핵심: AI drug discovery pipeline 전반 적용 확인 — transformative innovation 진행 중.

**AI Drug Discovery 종합 Review: 현재 한계와 기회[6]**
A. Seyhan et al.(Critical Reviews in Oncology/Hematology 2026)의 review는 AI drug discovery의 현재 상황을 종합한다. **AI 도구들**: AlphaFold·MoleR·PocketCrafter(scaffold-aware·3D molecular design), single-cell foundation models, large language models(LLMs). **Pipeline 적용 범위**: target identification·drug discovery·lead optimization·phenotypic screening·precision biology. **FDA approved AI devices**: numerous AI-enabled medical devices·software tools approved. **fully AI-discovered·designed drug**: 아직 marketing approval 0개. **AI-originated candidates**: 임상 개발 진행 중. **핵심 한계**: upstream data quality, downstream experimental validation constraints, regulatory·ethical frameworks 필요. 핵심: AI drug discovery의 현재 한계(fully AI-designed drug 0개)와 기회(AI-originated candidates 임상 진행) 동시 확인.

**8월 29일 Assessing Day vs 오늘의 차별점**: 8/29이 **Shamir JACS**으로 **AlphaFold3 covalent BTK inhibitor prospective discovery + co-crystallography validation**을, **Menon bioRxiv**으로 **AF3 vs DOCK3 complementarity · memorization**을 제시했다면, 오늘은 **Lin GEN Biotechnology**으로 **pharma 2026년 AI platforms 대규모 딜 급증**을(산업 동향 Mapping), **Ali Drug Development Research**으로 **AI drug discovery pipeline 전반 적용 종합**을(pipeline 적용), **Seyhan Critical Reviews**으로 **fully AI-designed drug 0개 한계 + AI-originated candidates 임상 진행**을(현실적 평가)를 동시에 제시한다. Assessing(8/29) → pharma AI 딜 급증 + pipeline 적용 + 한계/기회 Mapping(8/30).

실무 함의: AI drug discovery 전략팀에서 Lin et al.의 pharma AI 딜 동향을 분석하여 partnership strategy를 수립해야 한다.

*출처: [Lin Pharma AI 2026 Deals GEN Biotechnology 2026](https://consensus.app/papers/details/bc6e9ea3ec2e5025af707131d2fa0ca7/), [Ali AI Drug Discovery Drug Development Research 2026](https://consensus.app/papers/details/7a8e73e032a7505fa67758810d3c03ec/), [Seyhan AI Drug Discovery Critical Reviews Oncology Hematology 2026](https://consensus.app/papers/details/0e6b685fd2be56fb8a814115873a73d5/)*

*#AI #DrugDiscovery #Lin #GENG Biotechnology #PharmaAI #2026Deals #Ali #DrugDevelopmentResearch #Pipeline #Seyhan #CriticalReviews #OncologyHematology #FullyAIDesignedDrug #Zero #AIOriginated #Clinical #MappingDay #AlphaFold #MoleR #PocketCrafter #FoundationModel #LLM #TargetIdentification #LeadOptimization #Toxicity #Pharmacokinetics #PersonalizedMedicine #Regulatory #DataQuality #Collaboration #Pharma #Partnership #Investment #Innovation*

---

### Post 4: 단백질 AI — AlphaFold Whole Proteome Quality Assessment와 Adversarial Invariance 문제

🔬 **[단백질 AI] AlphaFold predictions on whole proteomes quality assessment로 pLDDT 한계 확인 + arity map으로 well-folded/IDP/implausible predictions 3개 population 분류(Sarti Bioinformatics Advances 2026) + AlphaFold3가 40% residues invariance + 10% deletions invariance 확인 — fold-switching proteins也不例外(Feldman Computational Structural Biotechnology 2026) + AlphaFold2-Multimer/3로 oligomeric states prediction의 utility와 limitations 동시 확인(Lin Protein Science 2026): AlphaFold의 현실적 한계 Mapping 동시 확인**

단백질 AI의**현실적 한계 Mapping**이 출처에서 동시에 확인된다[7][8][9].

**AlphaFold Whole Proteome Quality Assessment: pLDDT 한계와 Arity Map[7]**
Edoardo Sarti et al.(Bioinformatics Advances 2026)의 연구는 AlphaFold predictions의 whole proteome quality assessment를 수행한다. **pLDDT의 한계**: model organisms에서 30-40% residues가 low-confidence pLDDT range에 해당. **물리적으로 불가능한 구조**: pLDDT가 때때로 physically implausible structures를 flag하지 못함. **Arity map 도입**: histogram of per-residue neighbor counts로 3개 population 분류 — (1) well-folded proteins, (2) intrinsically disordered proteins(IDPs), (3) physically implausible predictions. **ABSRAT**: Arity-Based STRuctural Arrangement Quality assessmenT — per-residue scoring function으로 hallucinated helices·distorted beta strands 감지. **핵심**: AlphaFold whole proteome quality assessment로 pLDDT 한계 확인 + arity map으로 3개 population 분류 — 단백질 AI quality assessment의 새로운 기준.

**AlphaFold3 Adversarial Mutations: 40% Residues Invariance 확인[8]**
J. Feldman et al.(Computational and Structural Biotechnology Journal 2026)의 연구는 AlphaFold3의 adversarial evaluation을 수행한다. **40% residues invariance**: 점돌연변이와 deletion突变에 대해 predicted structures가 invariance 유지 — fold-switching proteins也不例外. **10% deletions invariance**: 10% deletions에도 invariance. **단백질의 종류 무관**: small monomeric proteins도 마찬가지 — AlphaFold가 best performance를 발휘해야 할 category에서도. **Confidence metrics 문제**: most accurate structure를 35% 경우에만 선택. **Training-set template 의존**: confidence metrics가 training-set template의 structural quality와 상관관계. **ESMFold 비교**: ESMFold가 더 큰_mutational sensitivity 보이나 여전히 imperfect. **핵심**: AlphaFold3 adversarial mutations에서 40% residues invariance 확인 — memorization vs biophysical reasoning 문제 재확인.

**AlphaFold Oligomeric State Prediction: Utility와 Limitations 동시 확인[9]**
Yiechang Lin et al.(Protein Science 2026)의 연구는 AlphaFold2-Multimer와 AlphaFold3의 oligomeric state prediction을 평가한다. **4,700개 proteins 이상 평가**: AlphaFold2-Multimer가 대부분의 경우 reliable oligomeric state predictions 제공. **한계**: training set에서 close structural representatives가 없는 proteins에서 accuracy 제한. ** membrane proteins 포함**: soluble·membrane proteins 모두 평가. **실용적 guidance**: computational cost 최소화, challenging cases identification, experimental data가 없는 proteins에 적용 guidance. **핵심**: AlphaFold oligomeric state prediction의 utility와 limitations 동시 확인 — 단백질 AI의 적용 범위 한계 Mapping.

**8월 29일 Assessing Day vs 오늘의 차별점**: 8/29이 **Menon bioRxiv**으로 **AlphaFold3 vs DOCK3 complementarity · memorization**을, **Shamir JACS**으로 **AF3 covalent BTK inhibitor prospective validation**을, **Kinde Results Engineering**으로 **AlphaFold drug discovery 전체 영역 종합**을 제시했다면, 오늘은 **Sarti Bioinformatics Advances**으로 **whole proteome quality assessment로 pLDDT 한계 + arity map 3개 population 분류**를(quality assessment 한계 Mapping), **Feldman Computational Structural Biotechnology**으로 **adversarial mutations에서 40% invariance 정량 확인**을(biophysics 한계 Mapping), **Lin Protein Science**으로 **oligomeric state prediction의 utility/limitations 동시 확인**을(적용 범위 Mapping)을 동시에 제시한다. Assessing(8/29) → AlphaFold의 quality assessment 한계 + adversarial invariance 한계 + oligomeric prediction 적용 범위 Mapping(8/30).

실무 함의: 전산단백질 설계팀에서 Feldman et al.의 adversarial invariance 분석을 참조하여 AlphaFold 적용 시 generalization limitation을 반드시 고려해야 한다.

*출처: [Sarti AlphaFold Quality Assessment Bioinformatics Advances 2026](https://consensus.app/papers/details/97d0d09907ed559998afbc1980b532f3/), [Feldman AlphaFold Adversarial Computational Structural Biotechnology 2026](https://consensus.app/papers/details/68f35194db525fb68a7bc8c10546001b/), [Lin AlphaFold Oligomeric States Protein Science 2026](https://consensus.app/papers/details/cff9b03d33d85cdf8d04f175b025a42c/)*

*#단백질AI #AlphaFold #Sarti #BioinformaticsAdvances #QualityAssessment #pLDDT #ArityMap #IDP #WellFolded #Feldman #ComputationalStructuralBiotechnology #Adversarial #Invariance #40Percent #FoldSwitching #Memorization #BiophysicalReasoning #ESMFold #Lin #ProteinScience #OligomericState #Multimer #Membrane #TrainingSet #Generalization #MappingDay #ConfidenceMetrics #ABSRAT #HallucinatedHelices #BetaStrands #ProteinStructure #Prediction #QualityControl #Benchmark*

---

### Post 5: Single-Cell Multi-Omics Foundation Model — VCBench 7 Dimensions와 Unified Benchmark Framework

🬬 **[Single-Cell Multi-Omics Foundation Model] VCBench가 7개 capability dimensions에서 single-cell foundation models의_virtual cell`_capability 평가 — baselines match or exceed every FM on 4/5 scored dimensions 확인(Weidener bioRxiv 2026) + unified framework가 13개 foundation models + 50개 datasets에서 systematic benchmark 수행(Hou bioRxiv 2026) + SCMBench에서 FMs << DMs + lightweight adaptation strategy 제시(Wang Nature Communications 2026): foundation model의 현실적 평가와 benchmark framework의 통합 Mapping 동시 확인**

Single-Cell Foundation Model의**현실적 평가와 benchmark framework 통합**이 출처에서 동시에 확인된다[10][11][12].

**VCBench: 7 Capability Dimensions에서 Baselines ≥ FMs 확인[10]**
L. Weidener et al.(bioRxiv 2026)의 연구는 VCBench 벤치마크를 제시한다. **7개 capability dimensions**: perturbation response prediction·cross-species universality·GRN inference·modality integration·temporal dynamics·multi-scale integration·in silico experimentation. **5개 foundation models 평가**: Geneformer·scGPT·UCE·TranscriptFormer·Arc State. **핵심 발견 1**: baselines match or exceed every FM on 4 of 5 scored dimensions — linear·nearest-neighbor baselines가 foundation models 능가. **핵심 발견 2**: TranscriptFormer만 cross-modal RNA-to-protein prediction에서 baseline 초과(53% Pearson improvement). **핵심 발견 3**: no FM이 complete cell-level training manifest.publish — contamination undetectable. **Contamination Reporting Schema**: VCBench의 기여. **핵심**: VCBench로 4/5 dimensions에서 baselines ≥ FMs 확인 — foundation model의 현실적 한계 Mapping.

**Unified Benchmark Framework: 13 Foundation Models · 50+ Datasets[11]**
Siyu Hou et al.(bioRxiv 2026)의 연구는 unified framework로 13개 foundation models를 benchmark한다. **소프트웨어 환경 통일**: software environments harmonization으로 manual configuration eliminated. **50개 datasets 이상**: zero-shot·few-shot·fine-tuning settings에서 평가. **핵심 발견 1**: pretrained embeddings가 biologically meaningful structure 포착 — low-label·transfer-learning scenarios에서 advantages. **핵심 발견 2**: classical PCA approach remains competitive or even preferable in some scenarios. **Best practices 제공**: technical barriers 낮추고 transparent·reproducible standard 확립. **핵심**: unified framework로 13개 foundation models + 50+ datasets systematic benchmark — benchmark의 통합적 수행.

**SCMBench: FMs << DMs + Lightweight Adaptation Strategy[12]**
Yixuan Wang et al.(Nature Communications 2026)의 SCMBench는 23개 방법 평가로 FMs << DMs 확인한다. **4개 평가 영역**: integration accuracy·biomarker detection·trajectory inference·batch effect correction. **FMs << DMs**: 모든 영역에서 foundation models가 domain-specific models보다 저조. **Lightweight adaptation strategy**: FMs strengths 유지하면서 DMs 수준 performance 향상. **FMs 강점**: versatility·robustness는 여전히 유지. 핵심: SCMBench benchmark로 foundation model의 현실적 한계 정량 확인 + lightweight adaptation solution 제시 — foundation model 한계의 Mapping.

**8월 27일 Converging Day vs 오늘의 차별점**: 8/27이 **Wang Nature Communications**으로 **SCMBench foundation models << domain-specific models**를, **Wu Genome Biology**으로 **no single scFM consistently outperforms**를 제시했다면, 오늘은 **Weidener bioRxiv**으로 **VCBench 7 dimensions에서 baselines ≥ FMs on 4/5 dimensions**를(새 benchmark), **Hou bioRxiv**으로 **unified framework 13 FMs + 50+ datasets**를(새 framework), **Wang Nature Communications**으로 **SCMBench의 lightweight adaptation strategy**를(integrated solution)를 동시에 제시한다. Converging(8/27) → 새로운 benchmark framework(VCBench) + unified framework + adaptation solution Mapping(8/30).

실무 함의: Single-cell 분석팀에서 Weidener et al.의 VCBench 결과를 분석하여 foundation model 도입 시 baselines 대비 성능 향상을 반드시 확인해야 한다.

*출처: [Weidener VCBench bioRxiv 2026](https://consensus.app/papers/details/48f053efc3485e6589bcd85cc6009842/), [Hou Unified Benchmark Framework bioRxiv 2026](https://consensus.app/papers/details/a7b39f95795558cc982213ff8dd7deb8/), [Wang SCMBench Foundation Models Nature Communications 2026](https://consensus.app/papers/details/24b273d01c1f508c9a20a5a981ac2e3c/)*

*#SingleCell #FoundationModel #VCBench #Weidener #VirtualCell #7Dimensions #Hou #UnifiedFramework #13Models #50Datasets #Wang #SCMBench #NatureCommunications #Benchmark #FoundationModel #DomainSpecific #Adaptation #Lightweight #MappingDay #PerturbationResponse #CrossSpecies #GRN #ModalityIntegration #TemporalDynamics #Baselines #Linear #NearestNeighbor #Geneformer #scGPT #UCE #TranscriptFormer #ArcState #Contamination #BestPractices #Reproducibility #Standardization #TransferLearning #ZeroShot #FewShot #FineTuning #PCA #Integration #Biomarker #Trajectory #BatchCorrection*

---

### Post 6: FDA/규제 — FDA Gene Therapy Approvals 1998-2025 체계적 분석

🏥 **[FDA/규제] FDA gene therapy approvals 38개 CGT 1998-2025 체계적 분석으로 규제 evolution과 future directions 도출(Oo Journal Clinical Pharmacology 2026) + FDA CGT approvals 44개 May 2025까지 확인 + 86.8% orphan designation · 92.1% expedited pathway 사용 확인(Shahzad Clinical Pharmacology Therapeutics 2026) + 14개 pediatric use CGT approvals 확인(Kunanayagam Pediatric Research 2026): FDA gene therapy regulatory evolution의 Mapping 동시 확인**

FDA 규제 체계의**Gene Therapy Regulatory Evolution Mapping**이 출처에서 동시에 확인된다[13][14][15].

**FDA Gene Therapy Approvals 1998-2025: 규제 Evolution과 Future Directions[13]**
C. Oo et al.(Journal of Clinical Pharmacology 2026)의 연구는 FDA gene therapy approvals의 1998-2025 역사를 체계적으로 분석한다. **1998-2025 Gene therapy approvals**: slow start → recent marked acceleration. **RNA-based agents·viral/non-viral in vivo platforms·ex vivo genetically modified cell therapies**: 최근 acceleration의 주요 영역. **FDA platform-aligned·risk-based initiatives**: plausible mechanism framework·CMC flexibility initiative·advanced manufacturing technologies program. **과학적 도전**: high upfront costs·manufacturing complexity·payer constraints. **AI/ML-enabled analytics**: dose selection·safety evaluation·durability prediction에 적용. **핵심**: FDA gene therapy approvals 1998-2025 체계적 분석으로 규제 evolution과 future directions 도출 — regulatory evolution Mapping.

**FDA CGT Approvals 44개 May 2025까지: Orphan · Expedited · Postmarketing[14]**
Mahnum Shahzad et al.(Clinical Pharmacology & Therapeutics 2026)의 연구는 FDA CGT approvals의 premarketing·postmarketing evidence를 분석한다. **44개 CGTs approved May 2025까지**: 38개 through December 31, 2025. **86.8% orphan designation**: 대부분 orphan 질환 대상. **92.1% expedited pathway 사용**: rapid approval 위한. **1.6개 pivotal studies 필요**: mean number, 73.7% single study 기반 approval. **44.7% surrogate endpoints에만 의존**: clinical endpoints 아닌. **Postmarketing 공통**: 73.4% at least one postmarketing requirement/commitment. **임상 efficacy endpoint 문제**: 17.9%만 primary clinical efficacy endpoint 포함. **핵심**: FDA CGT approvals 44개 분석으로 규제 체계의 expedited approval과 postmarketing evidence 문제 Mapping.

**Pediatric CGT Approvals: 14개 pediatric use + PK/PD 고려사항[15]**
S. Kunanayagam et al.(Pediatric Research 2026)의 review는 pediatric CGT approvals의 pharmacokinetic과 dosing considerations를 분석한다. **44개 CGTs May 2025까지**: 이 중 14개가 pediatric use. **In vivo gene therapy 개발 고려사항**: vector shedding·biodistribution. **Ex vivo gene therapy**: traditional PK principles 적용 불가 — cellular kinetics(expansion·persistence)가 관련. **CAR-T·TILs**: cellular kinetics 중요. **Preclinical models 의존**: safety·efficacy·dose selection을 위한 preclinical data 의존. **Rare disease·pediatric challenges**: small patient populations으로 인한 추가 도전. **핵심**: pediatric CGT approvals 14개 + unique PK/PD 고려사항 Mapping.

**8월 29일 Assessing Day vs 오늘의 차별점**: 8/29이 **Wu Nature Medicine**으로 **FDA AI devices 130개 분석에서 high-quality trials 3.2%**를, **Hussain Heart**으로 **cardiology AI 277개 approved의 510(k) predicate creep 위험**을, **Warraich JAMA**으로 **nearly 1000 AI-enabled medical devices authorized**를 제시했다면, 오늘은 **Oo Journal Clinical Pharmacology**으로 **FDA gene therapy approvals 1998-2025의 전체 역사 분석**을(규제 evolution Mapping), **Shahzad Clinical Pharmacology Therapeutics**으로 **CGT 44개 approvals의 orphan·expedited·postmarketing 패턴**을(승인 패턴 Mapping), **Kunanayagam Pediatric Research**으로 **pediatric CGT 14개 + unique PK/PD**를(的特殊 집단 Mapping)를 동시에 제시한다. Assessing(8/29) → gene therapy regulatory evolution + approval patterns + pediatric considerations Mapping(8/30).

실무 함의: 규제 전략팀에서 Oo et al.의 gene therapy regulatory evolution 분석을 참조하여 장기적 regulatory strategy를 수립해야 한다.

*출처: [Oo FDA Gene Therapy Approvals Journal Clinical Pharmacology 2026](https://consensus.app/papers/details/181658e93c825ea581a0664845a98d12/), [Shahzad FDA CGT Assessment Clinical Pharmacology Therapeutics 2026](https://consensus.app/papers/details/24705d5037e85bcb9d9863f1aabca040/), [Kunanayagam Pediatric CGT Pediatric Research 2026](https://consensus.app/papers/details/1c13548ad5be537a8451eed2c85ecd75/)*

*#FDA #GeneTherapy #Regulatory #Oo #JournalClinicalPharmacology #1998-2025 #Evolution #Shahzad #ClinicalPharmacologyTherapeutics #44CGTs #OrphanDesignation #ExpeditedPathway #SurrogateEndpoints #Postmarketing #Kunanayagam #PediatricResearch #14Pediatric #PKPD #MappingDay #CMC #AdvancedManufacturing #AI #ML #DoseSelection #Safety #Durability #VectorShedding #Biodistribution #CAR-T #TILs #CellularKinetics #RareDisease #Pediatric #GeneTherapy #CellTherapy #RegulatoryScience #ApprovalPathways*

---

### Post 7: Spatial Transcriptomics — SCLC 600K Cells Spatial Mapping과 Pan-Cancer Spatial Microenvironment

🔬 **[Spatial Transcriptomics] SCLC 105개 patients · 600K cells 이상에서 lymph node metastasis의 공간적 세포 생태계 매핑(Zhang Cell Reports Medicine 2026) + gastric cancer 4개 공간 영역에서 lymphocyte-aggregated region(LAR) 발현과 tertiary lymphoid structures(TLS) 확인(Gao Nature Communications 2026) + pan-cancer 12 cancer types · 373 samples · 56 LCPs · 13 niches로 tumor spatial microenvironment(TME) 매핑(Li Cell Reports Medicine 2026): 종양 미세환경의 대규모 공간적 Mapping 동시 확인**

Spatial transcriptomics의**대규모 공간적 Mapping**이 출처에서 동시에 확인된다[16][17][18].

**SCLC Spatial Mapping: 105 Patients · 600K+ Cells · Lymph Node Metastasis[16]**
Zicheng Zhang et al.(Cell Reports Medicine 2026)의 연구는 SCLC의 lymph node metastasis에서 spatial cellular ecosystem을 매핑한다. **CosMx Spatial Molecular Imager 사용**: 105개 primary·metastatic lymph node specimens · 75개 SCLC patients. **600K+ cells 이상**: comprehensive atlas 생성. **3개 LNM-enriched malignant subclusters 확인**: distinct metabolic·angiogenic programs. ** Immune exclusion features**: malignant subclusters가 공간적으로 immune exclusion과 상관관계. **Vascular-immune crosstalk**: endothelial cells가 malignant cells 회피 + cytotoxic T cells와 functional perivascular niches 형성. **Pan-immune hotspot(PIHs-1)**: 독립적 survival predictor로 확립. **핵심**: SCLC 600K+ cells spatial mapping으로 lymph node metastasis의 공간적 세포 생태계 매핑 — 종양 미세환경 Mapping.

**Gastric Cancer LAR: Lymphocyte-Aggregated Region과 TLS[17]**
Sen Gao et al.(Nature Communications 2026)의 연구는 gastric cancer의 spatial regions를 매핑한다. **4개 공간 영역 확인**: integrated single-cell·spatial transcriptomics. **Lymphocyte Aggregated Region(LAR)**: lymphocyte aggregates + tertiary lymphoid structures(TLS) 포함. **Naive T cell abundance와 T cell activation pathways**: LAR 내에서 상관관계. **Group A vs Group B**: PD1+CD27+CD8 T cells가 CD70+LAMP3+ dendritic cells와 근접한Group A, resting lymphocytes가 Group B. **Immunotherapy biomarkers**: gastric cancer immunotherapy biomarkers에 대한 통찰 제공. **핵심**: gastric cancer 4개 공간 영역에서 LAR과 TLS 확인 — 종양 면역 미세환경 Mapping.

**Pan-Cancer TME: 12 Cancer Types · 373 Samples · 56 LCPs · 13 Niches[18]**
Jiarong Li et al.(Cell Reports Medicine 2026)의 연구는 pan-cancer spatial transcriptomic analysis를 수행한다. **12 cancer types · 373 samples**: 체계적 분석. **56개 Local Cellular Programs(LCPs)**: tumor spatial microenvironment(TSME) 내에서. **13개 recurrent niches 확인**: niche-shared·niche-specific interactions. **Niche_4(Macrophages + Tumor cells)**: poor prognosis·immunotherapy resistance 상관관계. **Niche_11(Macrophages + Immune cells)**: better survival·treatment response 예측. **Tumor cells·macrophages의 위치 의존성**: gene expression이 위치에 따라 결정. **핵심**: pan-cancer 12 cancer types · 56 LCPs · 13 niches로 종양 미세환경 통합 Mapping.

**8월 29일 Assessing Day vs 오늘의 차별점**: 8/29이 **Norkin Cancer Research**으로 **CRC 40 PDOs + 16 PDO-tumor pairs Xenium spatial transcriptomics**를, **Pei Nature**으로 **pancreatic cancer 55 samples · 13 autopsies · 3 organs metastatic spatial mapping**을 제시했다면, 오늘은 **Zhang Cell Reports Medicine**으로 **SCLC 105 patients · 600K+ cells · 75 patients lymph node metastasis spatial ecosystem mapping**을(scale 확대), **Gao Nature Communications**으로 **gastric cancer 4개 공간 영역 LAR · TLS 구체적 매핑**을(specific region Mapping), **Li Cell Reports Medicine**으로 **pan-cancer 12 cancer types · 56 LCPs · 13 niches 통합 Mapping**을(integrated cross-cancer)을 동시에 제시한다. Assessing(8/29) → SCLC 대형 매핑 + gastric cancer specific region + pan-cancer 통합 Mapping(8/30).

실무 함의: 면역종양학팀에서 Li et al.의 pan-cancer niche 분석을 참조하여 niche-based patient selection criteria를 수립해야 한다.

*출처: [Zhang SCLC Spatial CosMx Cell Reports Medicine 2026](https://consensus.app/papers/details/dc8b02ddbfcb574fb92b18cf8aed1117/), [Gao Gastric Cancer LAR Nature Communications 2026](https://consensus.app/papers/details/46f3d71acd2e5b6e8816d6886aa485d9/), [Li Pan-Cancer TME Cell Reports Medicine 2026](https://consensus.app/papers/details/f420ab985e1257e18352cac9e0c425cf/)*

*#SpatialTranscriptomics #Zhang #SCLC #CosMx #600Kcells #LymphNodeMetastasis #Gao #GastricCancer #LAR #TLS #Li #PanCancer #12CancerTypes #373Samples #56LCPs #13Niches #TME #MappingDay #SpatialMapping #CellularEcosystem #Niche #PIHs1 #PanImmuneHotspot #EndothelialCells #CytotoxicTCells #Vascular #Macrophages #TumorCells #ImmuneExclusion #PD1 #CD27 #CD8T #CD70 #LAMP3 #DendriticCells #Immunotherapy #Biomarker #Prognosis #Resistance #Response #Metastasis #Metastatic #LNM #Primary #LymphNode #Organoid #PDO #SpatialHeterogeneity #TertiaryLymphoidStructures #LocalCellularProgram #RecurrentNiche #TumorMicroenvironment*

---

### Post 8: 오늘의 요약 — Mapping(매핑 Day) 핵심 정리

🧬 **오늘 6개 영역 핵심 요약:**

8/30(Mapping)의 핵심: Assessing(8/29)된 기술들이 **대규모 시공간적 매핑으로 데이터베이스화하고 통합 가능성을 재확인하는 날**이었다.

**1. Nanomaterial-Mediated CRISPR/Cas Delivery 확장(Wang Frontiers Bioengineering 2026)**: LNPs에서 vesicle-derived systems까지 확장 + ionizable LNPs로 93% protein knockdown(Yang MedComm 2026) + therapeutic applications 확장(Makhijani Discover Nano 2026) — CRISPR delivery의 nanomaterial 기반 확장[1][2][3].

**2. Pharma 2026 AI Deals 급증(Lin GEN Biotechnology 2026)**: 2026년 AI platforms 대규모 딜 급증 확인 + pipeline 전반 적용(Ali Drug Development Research 2026) + fully AI-designed drug 0개 + AI-originated candidates 임상 진행(Seyhan Critical Reviews 2026) — pharma AI 딜 급증과 현주소[4][5][6].

**3. AlphaFold Whole Proteome Quality Assessment(Sarti Bioinformatics Advances 2026)**: pLDDT 한계 확인 + arity map으로 3개 population 분류. 40% residues invariance adversarial mutation 확인(Feldman Computational Structural Biotechnology 2026) + oligomeric state prediction utility/limitations 동시 확인(Lin Protein Science 2026) — AlphaFold의 현실적 한계 Mapping[7][8][9].

**4. VCBench 7 Dimensions에서 Baselines ≥ FMs 4/5 확인(Weidener bioRxiv 2026)**: 새로운 benchmark framework + unified framework 13 FMs + 50+ datasets(Hou bioRxiv 2026) + SCMBench lightweight adaptation(Wang Nature Communications 2026) — foundation model benchmark의 통합적 수행[10][11][12].

**5. FDA Gene Therapy Approvals 1998-2025 체계적 분석(Oo Journal Clinical Pharmacology 2026)**: gene therapy regulatory evolution + CGT 44개 approvals orphan 86.8% · expedited 92.1%(Shahzad Clinical Pharmacology Therapeutics 2026) + pediatric CGT 14개(Kunanayagam Pediatric Research 2026) — regulatory evolution Mapping[13][14][15].

**6. SCLC 600K+ Cells Spatial Mapping(Zhang Cell Reports Medicine 2026)**: lymph node metastasis spatial ecosystem + gastric cancer LAR · TLS(Gao Nature Communications 2026) + pan-cancer 12 types · 56 LCPs · 13 niches(Li Cell Reports Medicine 2026) — 종양 미세환경 대규모 Mapping[16][17][18].

**8월 31일 전망**: Mapping된 대규모 데이터들이 다음 단계에서 **Integrating(통합)되어 cross-domain insights를 도출하는 단계**로 진입할 것으로 예상된다. 특히 pan-cancer niche analysis와 foundation model benchmark의 통합이 핵심 변수가 될 것으로 보인다.

*출처: [Wang Nanomaterial CRISPR Frontiers 2026](https://consensus.app/papers/details/f43692af0ccb54b684ef6714b3de8d82/), [Yang Genome Editing MedComm 2026](https://consensus.app/papers/details/87a151ebbb6759e48db9d0a028ff4597/), [Makhijani CRISPR Discover Nano 2026](https://consensus.app/papers/details/7acc7eabb0f05c51bccf0d568ec55751/), [Lin Pharma AI GEN Biotechnology 2026](https://consensus.app/papers/details/bc6e9ea3ec2e5025af707131d2fa0ca7/), [Ali AI Drug Discovery Drug Development Research 2026](https://consensus.app/papers/details/7a8e73e032a7505fa67758810d3c03ec/), [Seyhan AI Drug Discovery Critical Reviews 2026](https://consensus.app/papers/details/0e6b685fd2be56fb8a814115873a73d5/), [Sarti AlphaFold Quality Bioinformatics Advances 2026](https://consensus.app/papers/details/97d0d09907ed559998afbc1980b532f3/), [Feldman AlphaFold Adversarial Computational Structural Biotechnology 2026](https://consensus.app/papers/details/68f35194db525fb68a7bc8c10546001b/), [Lin AlphaFold Oligomeric Protein Science 2026](https://consensus.app/papers/details/cff9b03d33d85cdf8d04f175b025a42c/), [Weidener VCBench bioRxiv 2026](https://consensus.app/papers/details/48f053efc3485e6589bcd85cc6009842/), [Hou Unified Framework bioRxiv 2026](https://consensus.app/papers/details/a7b39f95795558cc982213ff8dd7deb8/), [Wang SCMBench Nature Communications 2026](https://consensus.app/papers/details/24b273d01c1f508c9a20a5a981ac2e3c/), [Oo FDA Gene Therapy Journal Clinical Pharmacology 2026](https://consensus.app/papers/details/181658e93c825ea581a0664845a98d12/), [Shahzad FDA CGT Clinical Pharmacology Therapeutics 2026](https://consensus.app/papers/details/24705d5037e85bcb9d9863f1aabca040/), [Kunanayagam Pediatric CGT Pediatric Research 2026](https://consensus.app/papers/details/1c13548ad5be537a8451eed2c85ecd75/), [Zhang SCLC Spatial Cell Reports Medicine 2026](https://consensus.app/papers/details/dc8b02ddbfcb574fb92b18cf8aed1117/), [Gao Gastric LAR Nature Communications 2026](https://consensus.app/papers/details/46f3d71acd2e5b6e8816d6886aa485d9/), [Li PanCancer TME Cell Reports Medicine 2026](https://consensus.app/papers/details/f420ab985e1257e18352cac9e0c425cf/)*

*#MappingDay #Summary #Wang #Nanomaterial #Yang #93PercentKnockdown #Makhijani #Lin #PharmaAI #Ali #Seyhan #FullyAIDesignedDrug #Zero #Sarti #AlphaFold #Quality #Feldman #Invariance #40Percent #Lin #Oligomeric #Weidener #VCBench #Hou #UnifiedFramework #Wang #SCMBench #Oo #FDA #Evolution #Shahzad #44CGTs #Kunanayagam #Pediatric #Zhang #SCLC #600K #Gao #LAR #TLS #Li #PanCancer #12Types #56LCPs #13Niches #Mapping #Assessing #Converging #SpatialTranscriptomics #TME #Niche #GeneTherapy #Regulatory #CRISPR #Delivery #Nanomaterial #AI #DrugDiscovery #FoundationModel #Benchmark #AlphaFold #Quality #ProteinStructure #VirtualCell #LargeScaleMapping*

---

## Sources

[1] [Wang et al. — Advances in nanomaterial-mediated CRISPR/Cas delivery: from lipid nanoparticles to vesicle-derived systems. Frontiers in Bioengineering and Biotechnology 2026](https://consensus.app/papers/details/f43692af0ccb54b684ef6714b3de8d82/)

[2] [Yang et al. — Delivery Systems for Therapeutic Genome Editing: Challenges, Innovations, and Future Perspectives. MedComm 2026](https://consensus.app/papers/details/87a151ebbb6759e48db9d0a028ff4597/)

[3] [Makhijani et al. — From mechanism to medicine: CRISPR‒Cas9 delivery strategies, therapeutic applications and translation challenges. Discover Nano 2026](https://consensus.app/papers/details/7acc7eabb0f05c51bccf0d568ec55751/)

[4] [Lin — Pharma Bets Big on AI Platforms with Flurry of 2026 Deals. GEN Biotechnology 2026](https://consensus.app/papers/details/bc6e9ea3ec2e5025af707131d2fa0ca7/)

[5] [Ali et al. — Artificial Intelligence in Drug Discovery and Development: Transforming Pharmaceutical Innovation. Drug Development Research 2026](https://consensus.app/papers/details/7a8e73e032a7505fa67758810d3c03ec/)

[6] [Seyhan et al. — AI in Drug Discovery and Development. Critical Reviews in Oncology/Hematology 2026](https://consensus.app/papers/details/0e6b685fd2be56fb8a814115873a73d5/)

[7] [Sarti et al. — Fold or flop: quality assessment of AlphaFold predictions on whole proteomes. Bioinformatics Advances 2026](https://consensus.app/papers/details/97d0d09907ed559998afbc1980b532f3/)

[8] [Feldman et al. — Adversarial Sequence Mutations in AlphaFold and ESMFold Reveal Nonphysical Structural Invariance, Confidence Failures, and Concerns for Protein Design. Computational and Structural Biotechnology Journal 2026](https://consensus.app/papers/details/68f35194db525fb68a7bc8c10546001b/)

[9] [Lin et al. — When can AlphaFold predict the oligomeric states of proteins? Protein Science 2026](https://consensus.app/papers/details/cff9b03d33d85cdf8d04f175b025a42c/)

[10] [Weidener et al. — VCBench: A Multi-Dimensional Benchmark for Single-Cell Foundation Models. bioRxiv 2026](https://consensus.app/papers/details/48f053efc3485e6589bcd85cc6009842/)

[11] [Hou et al. — A unified framework enables accessible deployment and comprehensive benchmarking of single-cell foundation models. bioRxiv 2026](https://consensus.app/papers/details/a7b39f95795558cc982213ff8dd7deb8/)

[12] [Wang et al. — SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration. Nature Communications 2026](https://consensus.app/papers/details/24b273d01c1f508c9a20a5a981ac2e3c/)

[13] [Oo et al. — FDA Gene Therapy Approvals (1998–2025): Current Status, Regulatory Evolution, and Future Directions. Journal of Clinical Pharmacology 2026](https://consensus.app/papers/details/181658e93c825ea581a0664845a98d12/)

[14] [Shahzad et al. — Assessment of United States Food and Drug Administration Approval of Cell and Gene Therapies. Clinical Pharmacology & Therapeutics 2026](https://consensus.app/papers/details/24705d5037e85bcb9d9863f1aabca040/)

[15] [Kunanayagam et al. — Clinical pharmacology insights from recent cell and gene therapy approvals relevant to pediatrics. Pediatric Research 2026](https://consensus.app/papers/details/1c13548ad5be537a8451eed2c85ecd75/)

[16] [Zhang et al. — Single-cell spatial transcriptomics reveals tumor microenvironment heterogeneity in primary and lymph node-metastatic small cell lung cancer. Cell Reports Medicine 2026](https://consensus.app/papers/details/dc8b02ddbfcb574fb92b18cf8aed1117/)

[17] [Gao et al. — A spatially resolved atlas of gastric cancer characterises a lymphocyte-aggregated region. Nature Communications 2026](https://consensus.app/papers/details/46f3d71acd2e5b6e8816d6886aa485d9/)

[18] [Li et al. — Pan-cancer analysis of spatial transcriptomics reveals heterogeneous tumor spatial microenvironment. Cell Reports Medicine 2026](https://consensus.app/papers/details/f420ab985e1257e18352cac9e0c425cf/)