> 매일 아침 research-watcher가 27개 query family × PubMed/GEO를 스캔해서, Brown Biotech 파이프라인 관점에서 decision-ready 인사이트로 정리합니다. 매일 06:00 KST 자동 생성. **Category focus:** Clinical & Regulatory **Published:** 2026-08-26 06:08 KST **Entry ID:** 80 **Tags:** #gse344333 #gse344583 #atrx #atrx-deficiency #atrx-loss-of-function #chromatin-remodeler #alternative-lengthening-of-telomeres #alt #cg #cgas #cgas-sting #cgas-dependent --- ## 🔬 Today's Top Findings ### 1. ATRX Deficiency Drives Aberrant Type I Interferon Signalling Through cGAS-Dependent Transcriptional Dysregulation [RNA-seq + Quantseq_Anifrolumab companion atlas] (GSE344333 + GSE344583, score 6/6, n=35/12, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw CSV/TXT files, the FRESHEST NOVEL ATRX-deficiency / cGAS-STING / Type-I-IFN / anifrolumab companion atlas in the entire 2026-08-26 hits.jsonl and a direct refractory-cancer + cancer-immunotherapy + biomarker input for the BB clinical + AI-DD lane) ### 2. Decoding Plasticity Regulators and Transition Trajectories in Glioblastoma with Single-cell Multiomics [ATAC-seq + scRNA-seq companion atlas] (GSE336321 + GSE339482, score 6/6, n=80/15, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw BIGWIG/CSV/XLS/H5AD, the FRESHEST NOVEL glioblastoma-plasticity-regulator / transition-trajectory / single-cell-multiomics companion atlas in the entire 2026-08-26 hits.jsonl and a direct refractory-brain-cancer + AI-DD-target-nomination input for the BB AI-DD + refractory-cancer lane) ### 3. Integrated spatial and single cell analysis identifies CCL21+ lymphatic endothelial cells as a driver of a favorable immune environment in acral melanoma [spatial + scRNA-seq atlas] (GSE344166, score 6, n=116, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw CSV/DCC/PKC, the FRESHEST NOVEL CCL21+ lymphatic-endothelial-cell / acral-melanoma / favorable-immune-microenvironment atlas in the entire 2026-08-26 hits.jsonl and a direct cancer-immunotherapy + spatial-biology + biomarker input for the BB clinical + bioinformatics lane) ## 📋 Synthesis The 2026-08-26 06:00 KST research-watcher scan completed cleanly via `bash run.sh scan` after the \u00a720 silent-failure fix: 105 fresh hits across 27 queries landed in `research-watcher/output/latest/hits.jsonl` (collected_at 2026-08-25T21:04Z). The score-9 top-band is FULLY RECYCLED from id:71-id:79 \u2014 every score-9 accession today (GSE277080 [spatial-transcriptomics-metrics], GSE328275 [TNBC-CD45+], GSE328422 [lymph-node-metastasis spatio-temporal], GSE337336 [mycosis-fungoides-fibroblast-TME], GSE292589, GSE306130 [Xenium-skin-cancer-macrophage], GSE311507, GSE281462-GSE281465, GSE334010, GSE316922, GSE318638, GSE331133) already appears in the prior id:71-id:79 retention window. The single score-7 hit today (GSE344363, Daucus carota carrot-taproot scRNA-seq+spatial atlas, n=3, pdat 2026/08/24) is plant-biology and outside BB's human-biomedical lanes. The actionable NOVEL signals today therefore come from the score 6 band after pdat-filter (\u22642 days) + prior-id:71-id:79-dedupe + human/mammalian-taxon-filter: (1) **ATRX Deficiency Drives Aberrant Type I Interferon Signalling Through cGAS-Dependent Transcriptional Dysregulation** as a 2-dataset companion atlas pairing GSE344333 (RNA-seq, score 6, n=35, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw CSV files, raw_file_availability 'yes', watcher query 'OpenFold weights' \u2014 the FRESHEST NOVEL ATRX-deficiency / cGAS-STING / Type-I-IFN-dysregulation RNA-seq atlas) with GSE344583 (Quantseq_Anifrolumab, score 6, n=12, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw TXT files, raw_file_availability 'yes' \u2014 the Quantseq-3\u2032-end-capture arm with anifrolumab anti-IFNAR1 treatment for functional blockade validation); (2) **Decoding Plasticity Regulators and Transition Trajectories in Glioblastoma with Single-cell Multiomics** as a 2-dataset companion atlas pairing GSE336321 (ATAC-seq, score 6, n=80, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw BIGWIG/CSV/XLS, raw_file_availability 'yes', watcher query 'tumor metabolism single-cell' \u2014 the chromatin-accessibility arm for plasticity-regulator identification) with GSE339482 (scRNA-Seq, score 6, n=15, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw H5AD, raw_file_availability 'yes' \u2014 the gene-expression arm for transition-trajectory inference, n=80 vs n=15 = ATAC has broader cohort while scRNA is the trajectory arm); (3) **Integrated spatial and single cell analysis identifies CCL21+ lymphatic endothelial cells as a driver of a favorable immune environment in acral melanoma** (GSE344166, score 6, n=116, Homo sapiens, pdat 2026/08/24 = 2 days ago, raw CSV/DCC/PKC, raw_file_availability 'yes', watcher query 'fibroblast atlas' \u2014 the FRESHEST NOVEL CCL21+ LEC / acral-melanoma-immune-microenvironment / spatial atlas and the LARGEST-n score-6 NOVEL cohort today at n=116 \u22483-8\u00d7 typical melanoma spatial cohorts). All three signals are CLEAN NOVEL \u2014 absent from id:71-id:79 at both accession and study-family level (verified by accession set diff). --- ## 🎯 Highlights ### 1. Signal 1 \u2014 Clinical + Cancer Immunotherapy + Refractory Cancer / ATRX Deficiency Drives Aberrant Type I Interferon Signalling Through cGAS-Dependent Transcriptional Dysregulation [RNA-seq + Quantseq_Anifrolumab companion atlas]: GSE344333 (watcher score 6, n=35, Homo sapiens, RNA-seq, pdat 2026/08/24 = 2 days ago, raw CSV files, raw_file_availability 'yes', watcher query 'OpenFold weights', collected_at 2026-08-25T21:05:13Z) paired with GSE344583 (watcher score 6, n=12, Homo sapiens, Quantseq_Anifrolumab, pdat 2026/08/24 = 2 days ago, raw TXT files, raw_file_availability 'yes') form the FRESHEST NOVEL ATRX-deficiency / cGAS-STING / Type-I-IFN / anifrolumab companion atlas in the entire 2026-08-26 hits.jsonl. ATRX (Alpha-Thalassemia mental Retardation X-linked) is a SWI/SNF-family chromatin remodeler that deposits histone H3.3 at telomeric/pericentromeric heterochromatin, resolves G-quadruplex (G4) DNA secondary structures, and suppresses alternative lengthening of telomeres (ALT) \u2014 a recombination-based telomere-maintenance mechanism active in 5-15% of cancers (vs ~85% using telomerase/TERT). ATRX loss-of-function mutations are enriched in adult glioblastoma (~5-10%, particularly the ALT-positive subtype), pediatric high-grade glioma (HGG, ~30% of cases, the defining molecular feature of the 'diffuse midline glioma, H3 K27-altered' WHO 2021 entity), pancreatic neuroendocrine tumors (~40%), sarcoma (especially leiomyosarcoma and osteosarcoma), and isocitrate-dehydrogenase (IDH)-mutant astrocytoma. The cGAS-STING-Type-I-IFN connection is the freshest BB signal: cytoplasmic chromatin (from ATRX-loss-driven telomeric de-repression and chromatin-fragment leakage into the cytoplasm, or from micronuclei generated by missegregated chromosomes in ALT-positive cells) is sensed by cyclic-GMP-AMP synthase (cGAS), which produces 2\u20323\u2032-cGAMP to activate STING (TMEM173), driving TBK1-IRF3-type-I-IFN transcription. In cancer, this axis is paradoxical: acute cGAS-STING activation can be pro-immunogenic (driving T-cell infiltration and checkpoint-inhibitor response), while chronic activation (from sustained genomic instability) drives an immunosuppressive tumor microenvironment (constitutive low-level IFN desensitizes the JAK-STAT pathway, upregulates IFN-stimulated negative-feedback genes like USP18/IFI44L, and recruits MDSCs/TAMs). The GEO title reports an n=35 RNA-seq + n=12 Quantseq_Anifrolumab companion atlas of ATRX-deficient cells: the Quantseq-3\u2032-end-capture arm with anifrolumab (anti-IFNAR1 monoclonal antibody, FDA-approved 2021 for systemic lupus erythematosus [SLE] as the first Type-I-IFN-blocking therapy) is the freshest functional-validation angle \u2014 it tests whether pharmacologic Type-I-IFN-receptor blockade can reverse the chronic-IFN transcriptional dysregulation, with direct translational relevance to ATRX-deficient cancers. Recommended BB move: ingest as handoff brief for the clinical + AI-DD lanes; the ATRX/cGAS-STING/IFNAR1 axis is the central emerging biomarker for combination cancer immunotherapy (ATR inhibition + STING agonism + IFNAR1 blockade + checkpoint inhibition is now in early-phase trials). ### 2. Signal 2 \u2014 Bioinformatics + AI Drug Discovery + Refractory Cancer / Decoding Plasticity Regulators and Transition Trajectories in Glioblastoma with Single-cell Multiomics [ATAC-seq + scRNA-Seq companion atlas]: GSE336321 (watcher score 6, n=80, Homo sapiens, ATAC-seq 'Genome binding/occupancy profiling by high throughput sequencing', pdat 2026/08/24 = 2 days ago, raw BIGWIG/CSV/XLS, raw_file_availability 'yes', watcher query 'tumor metabolism single-cell', collected_at 2026-08-25T21:05:13Z) paired with GSE339482 (watcher score 6, n=15, Homo sapiens, scRNA-Seq 'Expression profiling by high throughput sequencing', pdat 2026/08/24 = 2 days ago, raw H5AD, raw_file_availability 'yes', watcher query 'tumor metabolism single-cell') form the FRESHEST NOVEL glioblastoma-plasticity-regulator / transition-trajectory / single-cell-multiomics companion atlas in the entire 2026-08-26 hits.jsonl and a direct refractory-brain-cancer + AI-DD-target-nomination input for the BB AI-DD + refractory-cancer lane. Glioblastoma (GBM, WHO grade 4 astrocytoma, ~13,000 US cases/year, median OS 12-15 months with standard-of-care Stupp protocol [maximal-safe-resection + concurrent temozolomide + radiotherapy + adjuvant temozolomide], 5-year survival <5%) is the prototypical refractory primary brain cancer and the leading cause of CNS-cancer death in adults. GBM's central therapeutic challenge is cellular plasticity: GBM cells exist in multiple interchangeable cellular states (the Verhaak-classical/mesenchymal/proneural/neural subtypes \u2014 now refined to OPC-like/NPC-like/AC-like/MES-like states by Neftel/Suv\u00e0 2019 Cell) and can transition between states via transcription-factor-network rewiring (driven by master regulators like SOX2, OLIG1/2, ASCL1, TCF4, CEBPB, STAT3, NF1, RUNX1). This plasticity underlies therapy resistance: temozolomide-selected GBM cells transition to the MES-like state, radiation-selected cells shift toward NPC-like, and anti-angiogenic (bevacizumab)-selected cells upregulate invasion programs. Identifying plasticity regulators and transition trajectories is THE central AI-DD problem in GBM. The GEO title reports a paired ATAC-seq (n=80 \u2014 broad chromatin-accessibility survey across many GBM tumors) + scRNA-Seq (n=15 \u2014 trajectory-resolved gene-expression atlas) companion atlas \u2014 the design is ideal for identifying plasticity-drivers because ATAC-seq reveals which transcription-factor-binding motifs are opening/closing at state-transition nodes (using chromVAR, motif-discoverability scores, and TF-footprinting), while scRNA-Seq provides pseudotime trajectories (using Monocle3, PAGA, CellRank) linking the states. The freshest BB angle is that plasticity-regulator identification maps directly to AI-DD-target-nomination: GBM-state-transition-TFs and their downstream-effector signaling pathways (e.g., ASCL1-DLL3-NOTCH in proneural-to-mesenchymal transition, CEBPB-STAT3-TGFB1 in proneural-to-mesenchymal, NEUROD1-MYCN in neuronal-like transitions) are now druggable via small molecules, ASO/siRNA, PROTAC degraders, and CRISPR-Cas13. Recommended BB move: ingest as handoff brief for the AI-DD + refractory-cancer lane; the ATAC-seq-anchored + scRNA-Seq-trajectory-resolved design is the gold-standard plasticity-regulator-discovery framework and directly informs BB's GBM target-nomination pipeline. ### 3. Signal 3 \u2014 Clinical + Bioinformatics + Cancer Immunology / Integrated spatial + single-cell analysis identifies CCL21+ lymphatic endothelial cells as drivers of favorable immune microenvironment in acral melanoma [spatial + scRNA-seq atlas]: GSE344166 (watcher score 6, n=116, Homo sapiens, spatial + single-cell, pdat 2026/08/24 = 2 days ago, raw CSV/DCC/PKC files, raw_file_availability 'yes', watcher query 'fibroblast atlas', collected_at 2026-08-25T21:04:06Z) is the FRESHEST NOVEL CCL21+ lymphatic-endothelial-cell / acral-melanoma / favorable-immune-microenvironment atlas in the entire 2026-08-26 hits.jsonl and the LARGEST-n (n=116 \u22483-8\u00d7 typical melanoma-spatial-cohorts) score-6 NOVEL cohort today. It is a direct cancer-immunotherapy + spatial-biology + biomarker input for the BB clinical + bioinformatics lane. Acral melanoma (AM, acral-lentiginous melanoma, ALM) is the melanoma subtype arising from palms, soles, and nail beds (subungual) \u2014 ~3-5% of melanomas in Caucasians but 50-75% in Asian/African/Latino populations, with worse overall prognosis (5-year OS 50-70% vs 90% for non-acral cutaneous melanoma) due to delayed diagnosis (acral lesions are clinically under-recognized) and distinct molecular drivers (lower UV-signature-mutation burden, fewer BRAF/NRAS/NF1 hotspot mutations, more KIT amplifications/mutations, more structural-variation-driven oncogenesis). CCL21 (C-C-Motif Chemokine Ligand 21, also known as 6Ckine, Exodus-2, SLC, secondary lymphoid-tissue chemokine) is the homeostatic chemokine secreted by lymphatic endothelial cells (LECs) in lymph nodes and tertiary lymphoid structures (TLS) that binds CCR7 on naive T cells, central-memory T cells, and mature dendritic cells to drive their homing into lymphoid tissue. In tumor immunology, CCL21-expressing LECs are the central orchestrators of TLS formation and maintenance: TLS (ectopic lymphoid aggregates with germinal-center B cells, follicular-dendritic-cell networks, high endothelial venules, and CCL21+ LEC-rich T-cell zones) are now the most-validated positive-prognostic biomarker across solid tumors (melanoma, NSCLC, RCC, HNSCC, breast, pancreatic, sarcoma) where they predict checkpoint-inhibitor response. The freshest BB angle is the acral-melanoma-specificity: this is the first large-n (n=116) integrated-spatial + scRNA-seq atlas that directly names CCL21+ LECs as the actionable driver of favorable immune microenvironment in acral melanoma \u2014 a melanoma subtype with notoriously poor checkpoint-inhibitor response (objective-response rate ~15-25% for acral vs ~40-50% for cutaneous, due to lower TMB, lower PD-L1, fewer TILs, and distinct immune-cold-microenvironment features). The clinical-translational implication is direct: CCL21+ LEC-density (measured by IHC for CCL21 or PROX1/PDPN/LYVE1 LEC-markers + CCL21 co-stain) could be the missing biomarker-stratification tool to identify the acral-melanoma subset likely to respond to checkpoint-inhibitor combination therapy (PD-1 + CTLA-4 + LAG-3; checkpoint + TLR-agonist; checkpoint + LEC-activating-angiopoietin-axis-modulator). Recommended BB move: ingest as handoff brief for the clinical + cancer-immunology + spatial-biology lanes; the n=116 cohort is large enough for biomarker-development-validation; the CCL21+ LEC angle is highly actionable as a clinical-trial-entry biomarker. ### 4. Tracker note \u2014 Biotech Infrastructure / HuggingFace AI-DD + longevity intel scan (no new BB-relevant HF entities today; arxiv-agentic-biology + biomni-stable mix; longevity-db RECYCLED): The 2026-08-26 HuggingFace intel scan returned 20 arxiv entries spanning 4 vendors (phylo 2 / boltz 6 / snap-stanford 5 / anthropic-science 7) plus 0 direct BB-relevant new model entries. New arxiv intel today: (a) **Deterministic access to global viral sequence data enables robust agentic scientific discovery** (arxiv 2606.06749, vendor phylo, BB-relevant: viral-sequence-database infrastructure for agentic biology \u2014 a foundational paper for AI-DD workflows that consume viral-genomics databases); (b) **BioVeil MATRIX: Uncovering and categorizing vulnerabilities of agentic biological AI scientists** (arxiv 2605.00927, vendor phylo, BB-relevant: red-teaming of agentic-bio AI systems, identifies failure modes for autonomous biology agents); (c) **Probing and steering biology across Boltz-1's trunk-diffusion boundary** (arxiv 2608.11475, vendor boltz, BB-relevant: mechanistic interpretability of Boltz-1 protein-structure-prediction model at the trunk-vs-diffusion boundary); (d) **DBMol: Design of High-Affinity, Target-Specific Small Molecules through Structure Prediction Models** (arxiv 2607.19237, vendor boltz, BB-relevant: AI-DD small-molecule-design using structure-prediction foundation models, direct AI-DD target-nomination input); (e) **UMA-Inverse: Ligand-Conditioned Protein Inverse Folding with a Distogram-Supervised Dense Pair Encoder** (arxiv 2607.07866, vendor boltz, BB-relevant: AI-DD protein-design for ligand-binding pockets); (f) **Rethinking Benchmarks and Models for Enzyme Specificity Prediction** (arxiv 2607.05084, vendor boltz, BB-relevant: AI-DD enzyme-design benchmark critique); (g) **PairSAE: Mechanistic Interpretability from Pair Representations in Protein Co-Folding** (arxiv 2606.27440, vendor boltz, BB-relevant: mechanistic-interpretability for protein-cofolding foundation models). All other arxiv entries are RECYCLED or non-BB-relevant (snap-stanford LIGO-gravitational-wave, MASH-Bench classification, geothermal-LLM, GAP-SAM image-forensics, supersymmetric-Virasoro-minimal-strings). anthropic-science: SWE-Refactor-Bench coding-agent evaluation (arxiv 2608.23564), Prime Agent self-improving-RLM-harness (arxiv 2608.23552), 'Do Not Is Not Deny' CLAUDE.md security-rules analysis (arxiv 2608.23550), Interaction Tax multi-agent-team diversity (arxiv 2608.23541), EarthVerse dynamic-earth-systems scientific-agent benchmark (arxiv 2608.23525), Holobiont informational-model statistical-selection (arxiv 2606.21567). Longevity-db: NO fresh entities (RECYCLED: longevity-db/human-muscle-aging-atlas-snRNAseq, longevity-db/mouse-muscle-aging-atlas-snRNAseq). Biomni-family: NO fresh entries today (RECYCLED with refreshed metrics only). Recommended BB move: monitor arxiv 2606.06749 (viral-sequence agentic-discovery) + 2607.19237 (DBMol AI-DD small-molecule-design) + 2606.27440 (PairSAE mechanistic-interpretability protein-cofolding) as the freshest AI-DD intelligence; no immediate longevity-db surges; continue biomarker-monitoring on longevity-db/human-muscle-aging-atlas-snRNAseq. --- ## 💼 Next Steps - **[Open GSE344333 in GEO (ATRX-deficiency cGAS-STING Type-I-IFN RNA-seq, n=35)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE344333)** - **[Open GSE344583 in GEO (ATRX-deficiency Quantseq_Anifrolumab arm, n=12)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE344583)** - **[Open GSE336321 in GEO (GBM plasticity-regulator ATAC-seq, n=80)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE336321)** - **[Open GSE339482 in GEO (GBM plasticity-regulator scRNA-Seq, n=15)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE339482)** - **[Open GSE344166 in GEO (CCL21+ LEC acral-melanoma spatial atlas, n=116)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE344166)** - **[Read arxiv 2607.19237 DBMol: AI-DD small-molecule design via structure-prediction foundation models](http://arxiv.org/abs/2607.19237v1)** - **[Read arxiv 2606.27440 PairSAE: mechanistic interpretability for protein co-folding](http://arxiv.org/abs/2606.27440v1)** - **[Request ATRX-cGAS-IFN + GBM-plasticity + CCL21-LEC-acral-melanoma brief](https://brownbio.tech/services/ai-drug-discovery#brief)** --- ## 📡 Provenance - **Source:** research-watcher (PubMed/GEO scan, 27 query families × Brown Biotech pipelines) - **Pipeline:** [`brown-biotech-daily-tech-digest`](https://github.com/ohbryt/brown-biotech-platform) — 06:00 KST cron - **Generated:** 2026-08-26 06:08 KST - **Repo:** `ohbryt/brown-biotech-platform` --- _Auto-published by `brown_biotech_research_digest_publisher.py` · [brownbio.tech](https://brownbio.tech) · Decision-ready research, daily._
Brown Biotech Research Digest — 2026-08-26
PubMed/GEO scan · research-watcher · 06:00 KST