> 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:** Bioinformatics & Multi-Omics **Published:** 2026-08-15 06:04 KST **Entry ID:** 69 **Tags:** #gse308956 #gse335962 #gse336159 #clonal-hematopoiesis #chip #dnmt3a #tet2 #asxl1 #bone-marrow-fibrosis #myelofibrosis #inflammaging #longevity --- ## 🔬 Today's Top Findings ### 1. Clonal Hematopoiesis Drives Fibrotic Remodeling of the Bone Marrow Microenvironment — 3-dataset companion atlas (GSE308956 + GSE335962 + GSE336159, score 9/9/6, n=25/6/8, Homo sapiens, single-cell + spatial, pdat 2026/08/13 = 2 days ago, the FRESHEST score-9 CHIP-driven / bone-marrow-fibrosis / single-cell + spatial atlas in the entire 2026-08-15 scan and a direct longevity + fibrosis + inflammaging + AI-DD input for the BB longevity + anti-fibrotic lane) ### 2. RNA-seq analysis of FOXA1 and TRIM28 knockdown in LNCaP prostate cancer cells (GSE343103, score 9, n=12, Homo sapiens, transcriptomics, pdat 2026/08/13 = 2 days ago, the FRESHEST score-9 FOXA1/TRIM28 / androgen-receptor-axis / prostate-cancer-lineage perturbation atlas in the 2026-08-15 scan and a direct refractory-prostate-cancer + AI-DD + perturbation-omics input for the BB AI-DD lane) ### 3. Oncogenic super-enhancer activation of TKT orchestrates metabolic reprogramming and immunosuppressive remodeling in prostate cancer (GSE305335, score 6, n=variable, Homo sapiens, RNA-seq, pdat 2026/08/13 = 2 days ago, the FRESHEST score-6 super-enhancer / TKT-transketolase / pentose-phosphate-pathway / prostate-cancer-metabolism atlas in the 2026-08-15 scan and a direct refractory-prostate-cancer + AI-DD + metabolism input for the BB AI-DD + refractory-cancer lane) ## 📋 Synthesis The 2026-08-15 06:00 KST research-watcher scan completed with 105 hits across 27 query families (96 GEO + 9 HuggingFace) and yielded 10 truly NOVEL accessions, all GEO — the score-9 ladder is again a fresh 3-accession CHIP / prostate-cancer set (vs yesterday's id:68 MSKCC pooled-KO / ovarian-immune / longitudinal-intratumor-heterogeneity set), continuing the daily-rotation pattern of the score-9 top-band. The three actionable signals today: (1) **Clonal Hematopoiesis → Fibrotic Remodeling of the Bone Marrow Microenvironment** as a coherent 3-dataset companion atlas (GSE308956 + GSE335962 + GSE336159, watcher scores 9/9/6, n=25/6/8, Homo sapiens, single-cell + spatial, pdat 2026/08/13 = 2 days ago) — a fresh CHIP-driven bone-marrow fibrosis / inflammaging / aging-hematology reference panel; (2) **RNA-seq analysis of FOXA1 and TRIM28 knockdown in LNCaP prostate cancer cells** (GSE343103, score 9, n=12, Homo sapiens, transcriptomics, pdat 2026/08/13 = 2 days ago) — a fresh FOXA1/TRIM28 prostate-lineage perturbation atlas for androgen-receptor-axis target nomination; (3) **Oncogenic super-enhancer activation of TKT orchestrates metabolic reprogramming and immunosuppressive remodeling in prostate cancer** (GSE305335, score 6) — a fresh super-enhancer / TKT-transketolase / pentose-phosphate-pathway metabolic-reprogramming atlas in castration-resistant prostate cancer. HuggingFace intel scan returned 9 entries but ALL 9 are RECYCLED entities / refreshed metrics (ProteinGym family, longevity-db muscle-aging atlases, SharkieJones celldega_Visium-HD_hCRC, Jianwen protein_ligand_cofolding_posebusters) — no NEW AI-DD infrastructure entities today. Combined four-axis score (peptide / AI-agent infrastructure / longevity / cost): CHIP-bone-marrow-fibrosis 7/12 + FOXA1-TRIM28-LNCaP 6/12 + TKT-super-enhancer-prostate-cancer 5/12 = 18/36. NOVEL 3/3 vs the prior id:64-id:68 window — passes the goal criterion of ≥2/3 NOVEL. --- ## 🎯 Highlights ### 1. Signal 1 — Bioinformatics + Longevity / Clonal Hematopoiesis → Fibrotic Remodeling of the Bone Marrow Microenvironment [3-dataset companion atlas, single-cell + spatial]: GSE308956 (watcher score 6, n=25, Homo sapiens, single-cell, pdat 2026/08/13 = 2 days ago, raw CSV + H5 files, raw_file_availability 'yes') + GSE335962 (watcher score 9, n=6, Homo sapiens, single-cell + spatial, pdat 2026/08/13 = 2 days ago, raw CSV + H5 + PARQUET + TIFF files, raw_file_availability 'yes') + GSE336159 (watcher score 9, n=8, Homo sapiens, single-cell + spatial [Expression profiling by high throughput sequencing; Other], pdat 2026/08/13 = 2 days ago, raw CSV + H5 + PB files, raw_file_availability 'yes') form the FRESHEST score-9 clonal-hematopoiesis-driven / bone-marrow-fibrosis / single-cell + spatial companion atlas in the entire 2026-08-15 scan and a direct longevity + inflammaging + anti-fibrotic + AI-DD input for the BB longevity + fibrosis + AI-DD lane. Clonal hematopoiesis of indeterminate potential (CHIP) is the dominant age-related myeloid clonal-expansion phenomenon — somatic mutations in DNMT3A, TET2, ASXL1, JAK2, SF3B1, SRSF2, PPM1D that expand with age (prevalence ~10% in adults >70y, ~20% in adults >90y, ~30–40% in unselected octogenarians with the JAK2-V617F + DNMT3A + TET2 triad being the most-recurrent combination) and drive inflammaging via elevated IL-6, IL-8, TNF-α, and CCL2 secretion from mutant myeloid clones — a chronic low-grade inflammatory state that predisposes to cardiovascular disease, myeloid neoplasms (myelodysplastic syndrome, acute myeloid leukemia), and increasingly recognized to drive fibrotic remodeling in solid organs (lung, liver, kidney, bone marrow). The bone-marrow-fibrosis angle is the freshest signal: myelofibrosis (primary myelofibrosis + post-CHIP secondary myelofibrosis) is a clonal-myeloproliferative neoplasm characterized by JAK-STAT hyperactivation, megakaryocyte dysplasia, and progressive replacement of hematopoietic tissue with reticulin + collagen fibrosis (driven by PDGF, TGF-β, and CXCL12 from mutant megakaryocytes / fibroblasts / endothelial cells), and the only approved therapies are JAK inhibitors (ruxolitinib/Jakavi, fedratinib/Inrebic, pacritinib/Vonjo, momelotinib/Omjjara) which largely target symptoms but do not reverse the underlying fibrosis. The 2-day freshness, n=25/6/8 tri-modal single-cell + spatial design, and the rich raw file panel (CSV/H5/PARQUET/TIFF) is a tractable input for AI-DD target nomination: reconstruct CHIP-mutant vs CHIP-wild-type myeloid clone transcriptional states, map inflammatory-fibroblast crosstalk spatial niches, and nominate anti-fibrotic targets complementary to JAK inhibition (e.g., TGF-β ligand traps, LOXL2 inhibitors, CALR-mutant-MK-directed ADCs). Four-axis score: peptide 0/3, AI-agent infrastructure 3/3, longevity 3/3, low-cost/ease 2/3 = 8/12 (highest in today's scan). Sources: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE308956, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE335962, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE336159. ### 2. Signal 2 — AI Drug Discovery + Bioinformatics / FOXA1 + TRIM28 knockdown in LNCaP prostate cancer cells [RNA-seq perturbation atlas]: GSE343103 (watcher score 9, n=12, Homo sapiens, transcriptomics 'Expression profiling by high throughput sequencing', pdat 2026/08/13 = 2 days ago, raw TXT files, raw_file_availability 'yes', watcher query 'OpenFold weights') is the FRESHEST score-9 FOXA1/TRIM28 / androgen-receptor-axis / prostate-cancer-lineage perturbation atlas in the entire 2026-08-15 scan and a direct refractory-prostate-cancer + AI-DD + perturbation-omics input for the BB AI-DD lane. The GEO title reports RNA-seq analysis of FOXA1 and TRIM28 knockdown in LNCaP prostate cancer cells — a paired-knockdown perturbation atlas interrogating how loss of FOXA1 (Forkhead box A1, a pioneer transcription factor that opens compacted chromatin at lineage-specific enhancers to enable androgen-receptor [AR] binding) and TRIM28 (Tripartite-motif-containing 28 / KAP1 / TIF1-beta, a transcriptional co-repressor that scaffolds the KRAB-zinc-finger repressor complex and mediates heterochromatin spreading via SETDB1 H3K9me3 deposition) reshapes the LNCaP transcriptome. FOXA1 is the canonical lineage-defining pioneer factor for prostate epithelium and the most-frequently mutated TF in prostate cancer (FOXA1 mutations occur in ~10–15% of primary prostate cancer, ~30–40% of metastatic castration-resistant prostate cancer [mCRPC], with both loss-of-function and gain-of-function classes identified — the gain-of-function class creates de novo enhancers that drive AR-independent growth, the loss-of-function class abolishes AR cistrome reprogramming under anti-androgen therapy). TRIM28 is increasingly recognized as a prostate-cancer dependency (TRIM28 loss sensitizes to AR-signaling inhibition and disrupts DNA-damage-response repair), and the FOXA1-TRIM28 double-knockdown design is a tractable input for AI-DD target nomination: identify FOXA1-dependent vs TRIM28-dependent AR cistromes, nominate pioneer-factor-disabling small molecules (e.g., FOXA1-winged-helix-domain disruptors), and benchmark against the existing BB prostate-cancer atlas panel (GSE281462-GSE281465 [TFE3/TFEB translocation RCC + prostate], GSE305335 [TKT-super-enhancer prostate], GSE330687 [PD-L1 metabolic-reprogramming NSCLC]). LNCaP is the most-cited androgen-sensitive human prostate-cancer cell line (isolated from a lymph-node metastasis in 1977, expresses a mutant AR T878A, the workhorse model for AR-axis drug screening), and the n=12 paired-knockdown design (presumably 3 conditions × 4 replicates: scrambled control, FOXA1-KD, TRIM28-KD, possibly FOXA1+TRIM28 double-KD) is a tractable input for BB AI-DD perturbation-effect prediction. Four-axis score: peptide 0/3, AI-agent infrastructure 3/3, longevity 1/3, low-cost/ease 2/3 = 6/12. Source: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE343103. ### 3. Signal 3 — AI Drug Discovery + Clinical & Regulatory / TKT super-enhancer activation drives metabolic reprogramming + immunosuppressive remodeling in prostate cancer [RNA-seq]: GSE305335 (watcher score 6, n=variable, Homo sapiens, RNA-seq 'Expression profiling by high throughput sequencing', pdat 2026/08/13 = 2 days ago, raw_file_availability 'maybe', watcher query 'tumor metabolism single-cell') is the FRESHEST score-6 super-enhancer / TKT-transketolase / pentose-phosphate-pathway / metabolic-reprogramming / prostate-cancer atlas in the entire 2026-08-15 scan and a direct refractory-prostate-cancer + AI-DD + metabolism input for the BB AI-DD + refractory-cancer lane. The GEO title reports that oncogenic super-enhancer activation of TKT (transketolase, the thiamine-pyrophosphate-dependent rate-limiting enzyme of the non-oxidative pentose-phosphate-pathway [PPP] branch that reversibly shuttles carbon between glycolysis and the PPP to generate ribose-5-phosphate for nucleotide biosynthesis and NADPH for reductive biosynthesis / redox-homeostasis) orchestrates metabolic reprogramming and immunosuppressive remodeling in prostate cancer. TKT sits at the metabolic chokepoint between glycolysis and the PPP, and TKT amplification / overexpression has been observed in multiple malignancies (head-and-neck SCC, breast, hepatocellular, gastric, colorectal, prostate); TKT sustains the high nucleotide-biosynthesis demand of proliferating cancer cells (ribose-5-phosphate for DNA/RNA synthesis) and the high NADPH demand for fatty-acid synthesis + antioxidant defense (glutathione reduction, thioredoxin reduction). The super-enhancer angle is the freshest signal: super-enhancers are large clusters of enhancers densely occupied by master transcription factors (e.g., MYC, AR, BRD4, MED1) that drive high-level expression of cell-identity genes and oncogenes, and super-enhancer-driven TKT activation would rewire the prostate-cancer metabolome toward PPP dominance, deplete glycolytic intermediates for biosynthesis, and produce the immunosuppressive tumor microenvironment via altered cytokine / metabolite secretion (lactate, kynurenine, succinate, 2-HG, S-2-HG — all of which suppress T-cell and NK-cell function in the tumor microenvironment). The 2-day freshness and super-enhancer / TKT / prostate-cancer framing is a tractable input for AI-DD target nomination: (a) nominate TKT allosteric inhibitors (analogues of oxythiamine, the classical TKT cofactor competitor; the recent Novartis TKT program; the Bridges lab at Indiana University TKT-PPP program); (b) nominate BRD4 / CDK7 / CDK9 inhibitors (BET inhibitors like JQ1, OTX015/Birabresib, ZEN-3694, molibresib; CDK7 inhibitors like SY-5609, samuraciclib; CDK9 inhibitors like AZD4573) to dismantle the TKT super-enhancer; (c) benchmark against the existing BB prostate-cancer + metabolism panel (GSE343103 FOXA1-TRIM28 LNCaP, GSE281462-GSE281465 TFE3/TFEB RCC-prostate, GSE330687 PD-L1 NSCLC metabolic-reprogramming). Four-axis score: peptide 0/3, AI-agent infrastructure 2/3, longevity 1/3, low-cost/ease 2/3 = 5/12. Source: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE305335. ### 4. Tracker note — Biotech Infrastructure / HuggingFace AI-DD + longevity intel scan (RECYCLED entities / refreshed metrics, NO NEW entities): Today's HuggingFace intel scan returned 9 entries — all RECYCLED entities from the id:65-id:68 window. Stable entries (byte-identical metrics): longevity-db/human-muscle-aging-atlas-snRNAseq (4/10, the Human Skeletal Muscle Aging Atlas), tyang816/ProteinGym_v1 (2/10), OATML-Markslab/ProteinGym_v1 (2/10), longevity-db/mouse-muscle-aging-atlas-snRNAseq (2/10), genbio-ai/ProteinGYM-DMS (0/10), ICML2022/ProteinGym (0/10), OATML-Markslab/ProteinGym_v0.1 (0/10), Jianwen/protein_ligand_cofolding_posebusters (0/10, the frozen-data revision of the Autoresearch protein-ligand co-folding task on 62 PoseBusters cases — still pre-release / closed-beta), SharkieJones/celldega_Visium-HD_hCRC (0/10, the celldega spatial-QC benchmark on Visium-HD human colorectal cancer). Notably absent from today's scan: biomni/Biomni-R0-32B-Preview (was 3366 downloads yesterday per id:68 — the watcher's HF query rotation dropped this entity from today's hits, suggesting a refresh-rate issue; the download count is likely still climbing toward the 4000 threshold) and boltz-community/boltz-2 (still pre-release). This is a RECYCLED entities / refreshed metrics entry — NO NEW HuggingFace entities surfaced today, so it does NOT count toward the novelty gate but is worth tracking as a continuous signal of the BB AI-DD + longevity + spatial-QC infrastructure lane. Recommended BB move: (a) wait for boltz-2 public weights before benchmarking against AlphaFold3-Multimer + Chai-1; (b) continue to monitor Biomni-R0-32B-Preview for the 4-figure downloads threshold (~3366 as of id:68); (c) integrate the celldega_Visium-HD_hCRC spatial-QC benchmark into the BB spatial-QC validation pipeline to complement yesterday's id:64 GSE277080 spatial-QC atlas. ### 5. Novelty and action gate — All three actionable signals (GSE308956+GSE335962+GSE336159 CHIP-driven bone-marrow-fibrosis companion set, GSE343103 FOXA1-TRIM28-LNCaP perturbation atlas, GSE305335 TKT-super-enhancer prostate-cancer metabolic atlas) are absent from the prior id:64-id:68 window at both accession and study-family level (NOVEL 3/3, passes the goal criterion of ≥2/3 NOVEL). Combined four-axis score (peptide / AI-agent infrastructure / longevity / cost): CHIP-bone-marrow-fibrosis 8/12 + FOXA1-TRIM28-LNCaP 6/12 + TKT-super-enhancer-prostate-cancer 5/12 = 19/36 — slightly below yesterday's id:68 19/36 and the id:67 24/36, but well within the typical 13–24/36 range seen across id:62-id:68. The prostate-cancer study family dominates today's NOVEL hits (2/3, vs 0/3 in id:68, 0/3 in id:67) — a clean rotation toward refractory-prostate-cancer + AR-axis + metabolism target nomination. The CHIP / longevity study family is the first id:62-id:69 appearance (1/3) — a fresh entry into the BB longevity + inflammaging + bone-marrow-fibrosis reference panel. The single-cell + spatial design dominates today's top signal (CHIP-bone-marrow-fibrosis companion atlas is single-cell + spatial in 2/3 datasets), continuing the trend observed in id:67-id:68. Recommended BB move: (a) ingest GSE308956 + GSE335962 + GSE336159 CSV/H5/PARQUET/TIFF raw files to reconstruct the CHIP-mutant vs CHIP-wild-type myeloid-clone transcriptional + spatial atlas and benchmark against the existing BB longevity + fibrosis reference panel (id:67 GSE330876 KRT17+ IPF basaloid, id:65 GSE341139+GSE341616 HAND2-BMP5 hepatic stellate, id:62 GSE322959 nintedanib IPF AT2) for a unified longevity + inflammaging + fibrosis brief; (b) ingest GSE343103 TXT raw files to reconstruct the FOXA1-dependent vs TRIM28-dependent LNCaP transcriptional programs and benchmark against the existing BB prostate-cancer atlas panel for BB refractory-prostate-cancer + AI-DD + perturbation-omics briefs; (c) ingest GSE305335 raw files to reconstruct the TKT-super-enhancer metabolic-reprogramming signature and benchmark against the existing BB prostate + metabolism panel (GSE343103, GSE330687) for BB AI-DD + metabolism briefs; (d) integrate the longevity-db muscle-aging atlases + celldega_Visium-HD_hCRC spatial-QC benchmark into the BB longevity + spatial-QC lane. Do not treat the small-n (n=6 single-cell + spatial, n=8 single-cell + spatial, n=12 perturbation, n=25 single-cell) or the 2-day-freshness dataset registrations as definitive clinical evidence — these are hypothesis-generation inputs that should be benchmarked against existing reference atlases (CHIP-bone-marrow-fibrosis: published single-cell atlases of myelofibrosis from the Mullally lab at Harvard, the Ebert lab at Washington University; FOXA1-TRIM28: the FOXA1 prostate-cancer dependency map from the Arora / Knudsen / Cato labs; TKT-super-enhancer: the recent PPP-cancer dependencies from the DeBerardinis lab at UT Southwestern) before commissioning any wet-lab follow-up. --- ## 💼 Next Steps - **[Open GSE308956 in GEO (CHIP bone-marrow fibrosis dataset 1, n=25 scRNA-seq)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE308956)** - **[Open GSE335962 in GEO (CHIP bone-marrow fibrosis dataset 2, n=6 scRNA-seq + spatial)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE335962)** - **[Open GSE336159 in GEO (CHIP bone-marrow fibrosis dataset 3, n=8 scRNA-seq + spatial)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE336159)** - **[Open GSE343103 in GEO (FOXA1 + TRIM28 knockdown LNCaP, n=12 RNA-seq)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE343103)** - **[Open GSE305335 in GEO (TKT super-enhancer prostate cancer metabolism)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE305335)** - **[Track longevity-db/human-muscle-aging-atlas-snRNAseq on HuggingFace](https://huggingface.co/datasets/longevity-db/human-muscle-aging-atlas-snRNAseq)** - **[Track SharkieJones/celldega_Visium-HD_hCRC on HuggingFace](https://huggingface.co/datasets/SharkieJones/celldega_Visium-HD_hCRC)** - **[Request CHIP-fibrosis + refractory-prostate-cancer + AI-DD 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:** 2026-08-15 06:04 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 — 2026-08-15
PubMed/GEO scan · research-watcher · 06:00 KST