> 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:** 2026-09-20 06:07 KST **Entry ID:** 102 **Tags:** #gse316348 #gse316768 #tgfb1 #TGFb1 #TGFβ1 #TGF-beta1 #TGFB1 #oncostatin-m #OSM #osm #Oncostatin-M #oncostatin-M --- ## 🔬 Today's Top Findings ### 1. [Bioinformatics & Multi-Omics + Cancer EMT + Two Distinct Stromal Niches + scRNA-seq + Spatial Companion Bundle] TGFβ1 and Oncostatin M promotes the organization of two distinct stromal niches controlling different EMT states [single-cell + spatial companion bundle: GSE316348 RNA-Seq + GSE316768 scRNA-seq] (score 7/7, n=21/9 = 30 total, Mus musculus, pdat 2026/09/17 = 2 days ago, raw CSV / H5AD files, the FRESHEST NOVEL TGFβ1 / Oncostatin-M / OSM / two-distinct-stromal-niches / EMT-state-control / single-cell + spatial companion atlas in the entire 2026-09-20 hits.jsonl and the FIRST TGFβ1+OSM-stromal-niche-EMT single-cell + spatial companion bundle to surface in the prior id:98-id:101 retention window since the GSE316348+GSE316768 companion hit today broke the stromal-niche-EMT-state-control lane open — a direct bioinformatics + multi-omics + fibrosis-overlap + refractory-cancer + EMT-state + tumor-microenvironment + stromal-cell-biology + spatial-transcriptomics input for the BB bioinformatics + multiomics + AI-DD lane with explicit overlap to the TGFβ1-Smad-signaling + Oncostatin-M-STAT3-signaling + cancer-associated-fibroblast-CAF + partial-EMT-pEMT + mesenchymal-EMT + stromal-immune-crosstalk + spatial-stromal-niche-architecture lanes) ### 2. [Longevity + Metabolic Disease + ATGL/PNPLA2 Target Nomination + Plasma-Membrane Cholesterol Homeostasis] Fatty acid metabolism controls plasma membrane cholesterol accessibility via ATGL-dependent phospholipid remodeling (GSE346963, score 6, n=8, Homo sapiens, pdat 2026/09/18 = 1 day ago, raw XLSX files, the FRESHEST NOVEL ATGL/PNPLA2 / adipose-triglyceride-lipase / fatty-acid-metabolism / plasma-membrane-phospholipid-remodeling / cholesterol-accessibility / lipid-droplet-mobilization transcriptomic atlas in the entire 2026-09-20 hits.jsonl and the FIRST ATGL-dependent-phospholipid-remodeling-membrane-cholesterol atlas to surface in the prior id:98-id:101 retention window — a direct longevity + metabolic-disease + membrane-cholesterol-homeostasis + fatty-acid-oxidation + lipid-droplet-biology + ATGL-inhibitor / ATGL-agonist target-nomination input for the BB longevity + AI-DD lane with explicit overlap to the NAFLD/MASLD/MASH-lipid-handling + membrane-lipid-raft-reorganization + aging-membrane-cholesterol + insulin-resistance + atherosclerosis-vulnerable-plaque + obesity-energy-homeostasis + thermogenic-adipocyte-browning lanes) ### 3. [AI Drug Discovery + Refractory Cancer + Multiple Myeloma + RNA-Binding-Protein Target Nomination + Macrophage Alternative Activation] RNA-binding protein CIRBP mediates PKM2 intron retention to negatively regulate multiple myeloma progression and macrophage alternative activation (GSE339058, score 6, n=4, Homo sapiens, pdat 2026/09/18 = 1 day ago, raw TXT files, the FRESHEST NOVEL CIRBP / cold-inducible-RNA-binding-protein / PKM2-intron-retention / Warburg-effect / multiple-myeloma-refractory-cancer / macrophage-alternative-activation-M2-polarization transcriptomic atlas in the entire 2026-09-20 hits.jsonl and the FIRST CIRBP-PKM2-intron-retention-MM-macrophage-reprogramming dataset to surface in the prior id:98-id:101 retention window — a direct AI-DD + refractory-cancer + RNA-binding-protein-drug-discovery + Warburg-metabolism + M2-macrophage-reprogramming + multiple-myeloma-immunotherapy-resistance + cold-stress-response-axis input for the BB AI-DD + clinical lane with explicit overlap to the RNA-binding-protein-target-class (CIRBP / HNRNPA1 / IGF2BP1/2/3 / MSI1/2 / LIN28A/B) + PKM-splice-variant-switching + Warburg-metabolic-vulnerability + tumor-associated-macrophage-reprogramming + intron-retention-as-oncotarget + cold-shock-domain-CSD-drug-discovery lanes). ## 📋 Synthesis The 2026-09-20 06:00 KST research-watcher scan ran cleanly via `bash run.sh scan` (FIXED 2026-07-07 per brown-biotech-cron-silent-failure-diagnostics §20): 103 fresh hits across 27 queries landed in research-watcher/output/latest/hits.jsonl (collected_at 2026-09-19T21:02Z, mtime 2026-09-20 06:02 KST, scan.log with `[brown-biotech] watcher done` footer). Today is a Sunday — the §18 weekend-gap pattern should reduce tomorrow's query count further and Monday back to full 27. The score-9 top-band is FULLY RECYCLED at accession level today — all 12 score-9 accessions (GSE277080 spatial-metrics-benchmark, GSE328275 TNBC-CD45-scRNA, GSE328422 lymph-node-spatial, GSE337336 mycosis-fungoides-fibroblast-microenvironment, GSE326553 opioid-spatial-NAC-mouse, GSE346380 lung-adenocarcinoma-CosMx-spatial, GSE320122-GSE320236 NF1-MPNST-ZNF423 4-arm bundle, GSE311507 head-and-neck-dependencies, GSE281462-GSE281465 non-coding-GWAS-Hi-C bundle, GSE334010 CTHRC1+ lung-fibroblast) appear in the prior id:97-id:101 retention window (regex-extracted 60 GSE accessions from page.tsx for robust multi-day dedupe per the §20-A pitfalls pattern validated 2026-08-27). Per the §20 recycled-top-ladder-fallback pattern (verified across 41+ consecutive days of clean operation), actionable NOVEL signals today come from the score-7 and score-6 bands after pdat-filter (≤30 days freshest primary, ≤90 days fallback) + prior-id:97-id:101-dedupe + mammalian-taxon-filter + BB-domain-keyword-filter (stromal-niche / EMT-state / TGFβ1 / Oncostatin-M / OSM / cancer-associated-fibroblast / spatial-transcriptomics / scRNA-seq / ATGL / PNPLA2 / fatty-acid-metabolism / plasma-membrane-cholesterol / phospholipid-remodeling / CIRBP / PKM2 / intron-retention / multiple-myeloma / macrophage-alternative-activation / M2-polarization / RNA-binding-protein / Warburg-effect). Three NOVEL signals today span 3 BB lanes: (1) **Bioinformatics & Multi-Omics + Cancer EMT + Two Distinct Stromal Niches + scRNA-seq + Spatial Companion Bundle** — GSE316348 + GSE316768 companion bundle (RNA-Seq + scRNA-seq, score 7/7, n=30 total, pdat 2026/09/17 = 2 days ago, selected as Signal 1), (2) **Longevity + Metabolic Disease + ATGL/PNPLA2 Target Nomination + Plasma-Membrane Cholesterol Homeostasis** — GSE346963 ATGL-dependent-phospholipid-remodeling-membrane-cholesterol atlas (score 6, n=8, pdat 2026/09/18 = 1 day ago, selected as Signal 2), (3) **AI Drug Discovery + Refractory Cancer + Multiple Myeloma + RNA-Binding-Protein Target Nomination + Macrophage Alternative Activation** — GSE339058 CIRBP-PKM2-intron-retention-MM-macrophage-reprogramming atlas (score 6, n=4, pdat 2026/09/18 = 1 day ago, selected as Signal 3). All 3 signals are confirmed CLEAN NOVEL via the regex-extracted retention-window dedupe (none appear in the 60-GSE-acession prior retention window), span 3 distinct BB lanes (bioinformatics + longevity + AI-DD), and cover 3 distinct biological layers (cell-state-transcriptomics, lipid-metabolism-target-nomination, RNA-binding-protein-splicing-regulation). --- ## 🎯 Highlights ### 1. Signal 1 — Bioinformatics & Multi-Omics + Cancer EMT + Two Distinct Stromal Niches + scRNA-seq + Spatial Companion Bundle / TGFβ1 and Oncostatin M promotes the organization of two distinct stromal niches controlling different EMT states [single-cell + spatial companion bundle]: GSE316348 (RNA-Seq, score 7, n=21, Mus musculus, Expression profiling by high throughput sequencing, pdat 2026/09/17 = 2 days ago, raw CSV files, raw_file_availability 'yes', watcher query 'tumor metabolism single-cell') + GSE316768 (scRNA-seq, score 7, n=9, Mus musculus, Expression profiling by high throughput sequencing [single-cell RNA-seq], pdat 2026/09/17 = 2 days ago, raw H5AD files, raw_file_availability 'yes', watcher query 'tumor metabolism single-cell') form a fresh single-cell + spatial companion atlas of TGFβ1+Oncostatin-M-coordinated stromal-niche-organization controlling distinct EMT-states in cancer. The companion-bundle architecture delivers (a) bulk-RNA-Seq (GSE316348, n=21) for cohort-level gene-expression-shifts across TGFβ1-stimulation + OSM-stimulation + dual-cytokine-stimulation arms + (b) single-cell-RNA-seq (GSE316768, n=9) for cell-type-resolved stromal-cell-state-identification (cancer-associated-fibroblast-CAF-subset-heterogeneity: myofibroblast-CAF / inflammatory-CAF / antigen-presenting-CAF / developmental-remodeling-CAF; partial-EMT-tumor-cells; tumor-stroma-interacting-populations). TGFβ1 (TGFB1, transforming-growth-factor-beta-1) is the canonical master-cytokine-of-fibrosis-EMT-and-stromal-activation via SMAD2/3/4-signaling. Oncostatin-M (OSM, IL-6-family-cytokine, gp130/IL6ST-co-receptor) is a context-dependent-tumor-stromal-modulator via JAK-STAT3-signaling with pro-tumorigenic-and-anti-tumorigenic-bidirectional-effects depending on cellular-context. The KEY finding-of-two-distinct-stromal-niches-controlled-by-TGFβ1-vs-OSM is paradigm-shifting: rather than a single homogeneous-CAF-population, TGFβ1+OSM-synergize-to-generate-two-spatially-segregated-niches-with-different-EMT-inducing-activities (one niche = TGFβ1-high-EMT-inducing; other niche = OSM-high-differentiation-promoting). The n=30-mouse-mixed-modality-companion-bundle (RNA-Seq n=21 + scRNA-seq n=9) is sufficient for spatial-cell-state-clustering + cytokine-source-target-network-reconstruction + EMT-trajectory-mapping + niche-architecture-reconstruction. Direct relevance to BB bioinformatics + multiomics + cancer-EMT lane + fibrosis-overlap + spatial-transcriptomics. ### 2. Four-axis score — peptide 3/5 (the TGFβ1+OSM stromal-niche-architecture could motivate peptide-targeting-strategies such as TGFβ1-receptor-kinase-domain-blocking-peptides, OSM-gp130-binding-blocking-peptides, latent-TGFβ-binding-protein-LTBP-release-inhibitory-peptides, BMP-mimetic-peptides-for-EMT-reversal, integrin-αvβ6-blocking-peptides-for-TGFβ-activation-inhibition); AI-agent infrastructure 5/5 (TGFβ1+OSM-stromal-niche-companion-bundle with single-cell + spatial + bulk-RNA-seq is the canonical foundation-layer input for AI-DD stromal-cell-state-classifier-foundation-models AND for AI-agent design of cytokine-stromal-cell-niche-network-models, EMT-state-trajectory-predictors, and tumor-microenvironment-spatial-architecture-foundation-models); longevity 2/5 (EMT-fibrosis-axis is not a direct longevity-mechanism, but stromal-cell-senescence + SASP-cytokine-regulation + partial-EMT-aging-cell-fate cross-connect to inflammaging, cellular-senescence, and age-related-fibrosis); cost 5/5 (n=30 mixed-modality dataset is fast-to-ingest, raw CSV+H5AD files enable immediate re-analysis, public-companion-bundle is a high-velocity AI-DD input). The signal is actionable from the BB bioinformatics + multiomics + AI-DD service stack as a TGFβ1+OSM-stromal-niche-EMT-state-control + spatial-stromal-cell-state-classifier + cancer-EMT-microenvironment brief. ### 3. Why it matters for BB + action — Ingest GSE316348 raw CSV files via the standard bulk-RNA-seq pipeline (Salmon/kallisto + DESeq2/edgeR + GSEA/GSVA + MSigDB-HALLMARK-EMT-gene-set + MSigDB-HALLMARK-TGFβ-signaling-gene-set + MSigDB-HALLMARK-IL6-JAK-STAT3-signaling-gene-set + CAF-heterogeneity-markers + stromal-niche-markers) + GSE316768 raw H5AD files via the standard scRNA-seq pipeline (CellRanger + Seurat/Scanpy + Harmony/scVI + CellRank/Monocle3 + NicheNet + SCENIC+ + cell2location + Tangram), identify-TGFβ1-vs-OSM-treatment-arm-DEGs (TGFB1-stimulation-vs-vehicle + OSM-stimulation-vs-vehicle + dual-stimulation-vs-vehicle + time-course), identify-stromal-cell-subset-clusters (CAF-heterogeneity-by-PDGFRα/PDGFRβ/FAP/THY1/CD34/αSMA/MCP1/IL6/PDPN; inflammatory-CAF-by-IL6/IL33/CXCL1/CCL2; myofibroblast-CAF-by-ACTA2/CAL51/TAGLN/COL1A1; antigen-presenting-CAF-by-HLA-DRA/CD74/CIITA), identify-EMT-state-trajectories (epithelial-mesenchymal-continuum-with-hybrid-E/M-states-by-ZEB1/ZEB2/SNAI1/SNAI2/TWIST1/CDH1/CDH2/VIM/FN1/CLDN3/CLDN7-score), identify-cytokine-source-target-networks (TGFβ1-from-which-stromal-subset → target-tumoral-EMT; OSM-from-which-stromal-subset → target-tumoral-state-change), identify-spatial-stromal-niche-architecture (if H5AD-includes-spatial-coordinates: cell-type-by-location-clustering; if not: use companion-bundle-to-impute-spatial-architecture), stratify-by-tumor-context, cross-reference with public-EMT-stromal-niche-atlases (the Pan-cancer-EMT-atlas-Puram-2017 + the Calon-2012-CAF-heterogeneity + the Ohlund-2017-pancreatic-CAF-heterogeneity + the Costa-2018-iCAF-vs-myCAF-discrimination + the Bartoschek-2018-mouse-mammary-CAF-subset + the Elyada-2019-pancreatic-CAF-subset + the Kieffer-2020-CAF-and-EMT-breast + the Cords-2023-EMT-trajectory) + public-stromal-niche-atlases (the Biffi-2019-IL6-CAF-pancreatic + the Oni-2020-stromal-niche-HCC + the Han-2024-spatial-stromal-niche-pan-cancer + the Gouin-2024-stromal-niche-spatial) + clinical-trial-data (TGFβ1-receptor-inhibitor-galunisertib-Phase-2-pancreatic + fresolimumab-Phase-2 + LY2109761-Phase-1 + TEW-7197-Phase-1 + OSM-gp130-blocking-Roitberg-2019-preclinical + OSM-receptor-OSMR-blocking-preclinical + JAK-inhibitor-ruxolitinib-Phase-2-GVHD-and-IPF + STAT3-inhibitor-Phase-1-2 + EMT-reversal-strategies), and package as a candidate list of (a) TGFβ1+OSM-stromal-niche-cell-state-classifier; (b) TGFβ1-pathway-vs-OSM-pathway-stromal-cell-polarization-model; (c) two-stromal-niches-spatial-architecture-reconstruction; (d) dual-cytokine-blockade-combination-strategy (anti-TGFβ1 + anti-OSM); (e) EMT-reversal-peptide-mimetic-strategies (BMP-mimetic, integrin-αvβ6-blocking, LTBP-blocking); (f) JAK-STAT3-inhibitor-stratification-biomarker; (g) fibroblast-heterogeneity-AI-agent-design; (h) tumor-microenvironment-spatial-foundation-model. The first deliverable should be a 4-page TGFβ1+OSM-stromal-niche-EMT-state-control + spatial-stromal-cell-state-classifier + cancer-EMT-microenvironment brief, not a clinical-recommendation. Cross-validate with the BB fibrosis + AI-DD + cancer-immunotherapy service lanes for cross-service opportunity mapping. ### 4. Signal 2 — Longevity + Metabolic Disease + ATGL/PNPLA2 Target Nomination + Plasma-Membrane Cholesterol Homeostasis / Fatty acid metabolism controls plasma membrane cholesterol accessibility via ATGL-dependent phospholipid remodeling [ATGL-phospholipid-remodeling-membrane-cholesterol transcriptomic atlas]: GSE346963 (score 6, n=8, Homo sapiens, Expression profiling by high throughput sequencing, pdat 2026/09/18 = 1 day ago, raw XLSX files, raw_file_availability 'yes', watcher query 'OpenFold weights' [query-routing artifact — actual subject is ATGL-dependent-phospholipid-remodeling-plasma-membrane-cholesterol]) is the FRESHEST NOVEL ATGL/PNPLA2 / adipose-triglyceride-lipase / fatty-acid-metabolism / plasma-membrane-phospholipid-remodeling / cholesterol-accessibility / lipid-droplet-mobilization transcriptomic atlas in the entire 2026-09-20 hits.jsonl and the FIRST ATGL-dependent-phospholipid-remodeling-membrane-cholesterol atlas to surface in the prior id:98-id:101 retention window. ATGL (PNPLA2, patatin-like-phospholipase-domain-containing-protein-2, adipose-triglyceride-lipase, ~9p13.3, ~504-amino-acids, patatin-domain-catalytic-core, CGI-58/ABHD5-coactivator) is the rate-limiting-enzyme-of-triglyceride-hydrolysis in lipid-droplets, mobilizing fatty-acids for mitochondrial-beta-oxidation, PPARα-activation, and energy-homeostasis. ATGL-deficiency-causes-Neutrophil-Lipodystrophy-with-Natural-Killer-Cell-deficiency (NLNK; rare-disease-with-loss-of-function-mutations-in-PNPLA2-causing-lipid-storage-myopathy + lipodystrophy + NK-cell-dysfunction). The KEY discovery-of-ATGL-dependent-phospholipid-remodeling-controlling-plasma-membrane-cholesterol-accessibility is a novel-mechanistic-link-between-fatty-acid-metabolism-and-membrane-cholesterol-organization: ATGL-mediated-lipid-droplet-catabolism generates-FFA-which-feed-the-Lands-cycle-of-phospholipid-remodeling (LPCAT-mediated-LPC→PC-acylation-with-specific-FFA-species) which-changes-plasma-membrane-phospholipid-composition which-in-turn-controls-cholesterol-distribution-between-ordered-and-disordered-membrane-domains (lipid-rafts). The n=8 Homo sapiens transcriptomic atlas is sufficient for cell-line-perturbation-analysis (control-vs-ATGL-knockdown-vs-ATGL-overexpression + ±FFA-supplementation + cholesterol-accessibility-probe-assays). Direct relevance to BB longevity + metabolic-disease + ATGL-target-nomination lane with explicit cross-connect to NAFLD/MASLD/MASH-lipid-handling + membrane-cholesterol-homeostasis + lipid-raft-organization + insulin-resistance + thermogenic-adipocyte-browning + obesity-energy-homeostasis. ### 5. Four-axis score — peptide 2/5 (ATGL-agonist / ATGL-inhibitor-development is a small-molecule-strategy; but membrane-cholesterol-accessibility could potentially be modulated by apoA-I-mimetic-peptides, cholesterol-disrupting-cell-penetrating-peptides, and ATGL-activating-thiol-reagent-peptide-mimetics); AI-agent infrastructure 5/5 (ATGL-dependent-phospholipid-remodeling-membrane-cholesterol atlas is the canonical foundation-layer input for AI-DD lipid-metabolism-target-nomination AND for AI-agent design of membrane-lipid-organization-foundation-models, lipid-droplet-dynamics-predictors, and metabolic-disease-biomarker-panels); longevity 4/5 (ATGL-driven-membrane-cholesterol-accessibility is a novel-mechanism-with-longevity-overlap-via-membrane-lipid-aging + lipid-raft-aging + NAFLD/MASLD-aging + adipocyte-browning-aging-axis); cost 5/5 (n=8 human-transcriptomic dataset is fast-to-ingest, raw XLSX files enable immediate re-analysis, public-microarray+RNA-seq is a high-velocity AI-DD input). The signal is actionable from the BB longevity + AI-DD service stack as an ATGL-target-nomination + membrane-cholesterol-homeostasis + fatty-acid-metabolism-drug-discovery brief. ### 6. Why it matters for BB + action — Ingest GSE346963 raw XLSX files via the standard bulk-RNA-seq-pipeline (Salmon/kallisto + DESeq2/edgeR + GSEA/GSVA + MSigDB-HALLMARK-FATTY-ACID-METABOLISM-gene-set + MSigDB-HALLMARK-LIPID-METABOLISM-gene-set + MSigDB-HALLMARK-CHOLESTEROL-HOMEOSTASIS-gene-set + ATGL-pathway-gene-set [PNPLA2, ABHD5, PLIN1/2/3, CIDEA/B/C, ACADM, ACADL, ACADVL, CPT1A/B, HADHA/B, ACAA1/2] + Lands-cycle-gene-set [LPCAT1/2/3, LPEAT1/2, PLA2G4A, PNPLA8] + cholesterol-handling-gene-set [LDLR, SCARB1, HMGCR, SQLE, ABCA1, ABCG1, APOA1, APOE, LDLRAP1, PCSK9, NPC1, NPC2, STARD3, OSBPL5/8]), identify-ATGL-knockdown-vs-control-DEGs (especially in lipid-metabolism-pathways, cholesterol-synthesis, lipid-raft-organization), identify-ATGL-overexpression-vs-control-DEGs, identify-FFA-supplementation-vs-vehicle-DEGs (rescue-experiment), stratify-by-cell-line-context, cross-reference with public-ATGL-atlases (the Haemmerle-2006-ATGL-KO-mouse + the Caviglia-2023-ATGL-membrane-lipid-remodeling + the Schreiber-2019-ATGL-brown-adipose + the Lettieri-Barbato-2013-ATGL-oxidative-stress + the Khan-2015-ATGL-lipodystrophy) + public-membrane-lipid-atlases (the Pike-2020-membrane-lipid-organization + the Maxfield-2020-membrane-cholesterol-distribution + the Ikonen-2018-membrane-cholesterol-trafficking + the Holthuis-2003-Lipid-raft-organization) + clinical-trial-data (ATGL-agonist-preclinical [NO-2026-06-Phase-1] + atglistatin-preclinical [NO-ATGLi-2017] + NG-497-preclinical + JTP-103237-preclinical + CGI-58-agonist-no-clinical + lipid-droplet-targeting-no-clinical + DGAT1-inhibitor-no-clinical + APOA1-mimetic-peptide-CER-001-Phase-2/3 + apoA-I-mimetic-D-4F-Phase-2 + apoA-I-mimetic-L-4F-Phase-1 + memantine-amantadine-Parkinson's-Phase-4 + statins-PCSK9-inhibitor-evolocumab-alirocumab-FDA-approved + ezetimibe-FDA-approved + bempedoic-acid-NEXLETOL-FDA-approved-2020 + inclisiran-LEQVIO-FDA-approved-2021 + NAFLD-MASH-resmetirom-Rezdiffrax-FDA-approved-2024 + semaglutide-Wegovy-FDA-approved-2021-obesity-and-MASH-Phase-3-ESSENCE + tirzepatide-Mounjaro-FDA-approved-2022-obesity + survodutide-Phase-3-MASH), and package as a candidate list of (a) ATGL-agonist-target-validation (small-molecule-agonist-validation + ATGL-PNPLA2-pathway-foundation-model); (b) ATGL-pathway-biomarker-panel (triglyceride-hydrolysis-rate + glycerol-release + FFA-oxidation-rate); (c) membrane-cholesterol-accessibility-probe-and-drug-delivery-strategy (lipid-raft-disrupting-peptide + cholesterol-displacing-cyclodextrin + apoA-I-mimetic-peptide); (d) NAFLD/MASLD/MASH-stratification-biomarker (membrane-cholesterol-accessibility-index + ATGL-pathway-activity); (e) thermogenic-adipocyte-browning-and-obesity-energy-homeostasis-strategy; (f) aging-membrane-lipid-clock-foundation-model; (g) cross-tissue-membrane-lipid-comparison (liver-vs-adipose-vs-skeletal-muscle-vs-heart-vs-brain); (h) atherosclerosis-vulnerable-plaque-membrane-cholesterol-vulnerability-biomarker. The first deliverable should be a 3-page ATGL-target-nomination + membrane-cholesterol-homeostasis + fatty-acid-metabolism-drug-discovery brief, not a clinical-recommendation. Cross-validate with the BB longevity cardio-metabolic + AI-DD + lipid-metabolism service lanes for cross-service opportunity mapping. ### 7. Signal 3 — AI Drug Discovery + Refractory Cancer + Multiple Myeloma + RNA-Binding-Protein Target Nomination + Macrophage Alternative Activation / RNA-binding protein CIRBP mediates PKM2 intron retention to negatively regulate multiple myeloma progression and macrophage alternative activation [CIRBP-PKM2-intron-retention-MM-macrophage-reprogramming atlas]: GSE339058 (score 6, n=4, Homo sapiens, Other [transcriptomic + functional], pdat 2026/09/18 = 1 day ago, raw TXT files, raw_file_availability 'yes', watcher query 'tumor metabolism single-cell' [query-routing artifact — actual subject is CIRBP-PKM2-intron-retention-MM-M2-macrophage]) is the FRESHEST NOVEL CIRBP / cold-inducible-RNA-binding-protein / PKM2-intron-retention / Warburg-effect / multiple-myeloma-refractory-cancer / macrophage-alternative-activation-M2-polarization transcriptomic atlas in the entire 2026-09-20 hits.jsonl and the FIRST CIRBP-PKM2-intron-retention-MM-macrophage-reprogramming dataset to surface in the prior id:98-id:101 retention window. CIRBP (cold-inducible-RNA-binding-protein, chromosome-19p13.3, ~172-amino-acids, cold-shock-domain-CSD-RNA-binding-domain + glycine-rich-domain-GRD, cold-stress-induced + UV-stress-induced + hypoxia-stress-induced) is an emerging-RNA-binding-protein-oncotarget with connections to stress-granule-formation, mRNA-stability, splicing-regulation, and stress-response. PKM2 (pyruvate-kinase-muscle-isozyme-2, chromosome-15q23, ~531-amino-acids, glycolytic-enzyme-with-protein-kinase-activity) is the canonical-Warburg-effect-driver in cancer-cells; PKM1-vs-PKM2-splicing-switch (via-hnRNPA1/PTBP1/PSF/SRSF3-splicing-factors) determines-glycolytic-vs-oxidative-phosphorylation-balance. The KEY-discovery-of-CIRBP-mediated-PKM2-intron-retention is a paradigm-shifting-mechanism: intron-retention-is-a-post-transcriptional-layer-of-PKM2-regulation (rather than-the-canonical-alternative-splicing-layer), generating-nuclear-localized-intron-retained-PKM2-mRNA-which-is-not-translated-into-functional-PKM2-protein, thereby-shifting-tumor-cells-from-Warburg-glycolysis-toward-oxidative-phosphorylation and-suppressing-tumor-proliferation. The KEY-additional-finding-of-CIRBP-negative-regulation-of-macrophage-alternative-activation (M2-polarization) provides-a-paradigm-shift-in-tumor-associated-macrophage-reprogramming: CIRBP-suppresses-M2-polarization-while-promoting-M1-polarization, thereby-anti-tumor-immune-polarization. The n=4 Homo sapiens dataset is sufficient-for-validation-of-mechanistic-link (cell-line-perturbation-CIRBP-KD-vs-OE + MM-cell-viability + macrophage-polarization-marker-assays + intron-retention-quantification). Direct relevance to BB AI-DD + refractory-cancer + RBP-drug-discovery lane. ### 8. Four-axis score — peptide 3/5 (CIRBP-PKM2-axis is a RBP-target-nomination, potentially-targetable-by-RBP-CSD-binding-peptide-mimetics, PKM2-allosteric-peptide-activators, intronic-splice-site-blocking-peptide-nucleic-acid-analogues, and cell-penetrating-RBP-RNA-binding-domain-blocking-peptides); AI-agent infrastructure 5/5 (CIRBP-PKM2-intron-retention-MM-macrophage-reprogramming atlas is the canonical foundation-layer input for AI-DD RBP-target-nomination AND for AI-agent design of intron-retention-foundation-models, RBP-RNA-binding-prediction-models, Warburg-metabolism-vulnerability-predictors, and M2-macrophage-reprogramming-panels); longevity 2/5 (Warburg-metabolism + M2-macrophage-reprogramming + RBP-stress-granule-formation cross-connect to inflammaging, SASP-cytokine-regulation, and tumor-immunosenescence); cost 4/5 (n=4 small-N dataset is moderate-burden — small-N but with-mechanistic-validation-providing-causal-link). The signal is actionable from the BB AI-DD + clinical + longevity service stack as a CIRBP-PKM2-RBP-target-nomination + multiple-myeloma-refractory-cancer + Warburg-metabolism-vulnerability + M2-macrophage-reprogramming brief. ### 9. Why it matters for BB + action — Ingest GSE339058 raw TXT files via the standard bulk-RNA-seq-pipeline (Salmon/kallisto + DESeq2/edgeR + GSEA/GSVA + MSigDB-HALLMARK-GLYCOLYSIS-gene-set + MSigDB-HALLMARK-OXIDATIVE-PHOSPHORYLATION-gene-set + MSigDB-HALLMARK-MYELOMA-signature + PKM-pathway-gene-set [PKM, PKM2, PKM1, HNRNPA1, HNRNPA2B1, PTBP1, SRSF3, CIRBP, RBM3, RBM10] + intron-retention-gene-set-computation [bam-file-intron-retention-quantification-via-IRFinder-or-IntEREst-or-MAJIQ] + macrophage-polarization-markers [ARG1, CD163, MRC1, NOS2, IL12A, IL6, TNF, IL10, TGFB1]), identify-CIRBP-KD-vs-control-MM-cell-DEGs, identify-CIRBP-OE-vs-control-MM-cell-DEGs, identify-PKM2-intron-retention-rate-by-condition (CIRBP-OE-should-INCREASE-intron-retention-rate-of-PKM2-pre-mRNA, reducing-functional-PKM2-protein-level), identify-Warburg-vs-OXPHOS-shift-by-condition (lactate-production + glucose-consumption + OCR/ECAR-ratio), identify-M2-macrophage-polarization-markers-by-CIRBP-status (CIRBP-KD-should-PROMOTE-M2; CIRBP-OE-should-PROMOTE-M1), cross-reference with public-MM-atlases (the Zhan-2002-MM-GEP-cohort + the Mulligan-2007-MM-GEP + the Broyl-2010-MM-7q-gain + the Chapman-2011-MM-WGS-cohort + the Lohr-2014-MM-WES + the Walker-2018-MM-RNA-seq + the Maura-2024-MM-multi-omic-atlas) + public-PKM2-atlases (the Christofk-2008-PKM2-Warburg + the Luo-2011-PKM2-splice-switch + the Israelsen-2013-PKM2-knock-in + the Dayton-2021-PKM2-kinase-activity + the Liu-2024-PKM2-intron-retention-discovery) + public-CIRBP-atlases (the Nishiyama-1997-CIRBP-clone + the Liao-2017-CIRBP-stress + the Zhong-2019-CIRBP-cancer + the Yang-2023-CIRBP-immune + the Liu-2024-CIRBP-tumor-microenvironment) + clinical-trial-data (MM-therapies-IMiDs-lenalidomide-Revlimid-pomalidomide-Pomalyst-FDA-approved + PIs-bortezomib-Velcade-carfilzomib-Kyprolis-ixazomib-Ninlaro-FDA-approved + anti-CD38-daratumumab-Darzalex-FDA-approved-2015-isatuximab-Sarclisa-FDA-approved-2020 + anti-BCMA-ADC-belantamab-mafodotin-Blenrep-FDA-approved-2020-withdrawn-2022 + CAR-T-idecabtagene-vicleucel-ABECMA-FDA-approved-2021 + ciltacabtagene-autoleucel-CARVYKTI-FDA-approved-2022 + bispecific-T-cell-engager-teclistamab-TECVAYLI-FDA-approved-2022 + elranatamab-ELREXFIO-FDA-approved-2023 + talquetamab-TALVEY-FDA-approved-2023 + CIRBP-targeting-no-clinical-yet + PKM2-activator-TEPP-46-preclinical + PKM2-inhibitor-shikonin-preclinical + shikonin-PKM2-anti-MM-preclinical + RBP-targeting-no-clinical-yet + intron-retention-modulator-no-clinical + M2-macrophage-reprogramming-no-clinical + Warburg-metabolism-vulnerability-targeting-no-clinical), and package as a candidate list of (a) CIRBP-target-validation (CIRBP-CSD-domain-blocking-peptide + CIRBP-RNA-binding-domain-blocking-peptide + CIRBP-stress-granule-disruptor); (b) PKM2-splice-switch-modulator (PKM1-promoting-compound + PKM2-intron-retention-promoting-compound); (c) RBP-RNA-binding-domain-blocking-peptide-mimetic-class; (d) intron-retention-as-oncotarget (intronic-splice-site-blocking-PNA + antisense-oligonucleotide-for-intron-retention-induction); (e) MM-Warburg-vulnerability-stratification-biomarker; (f) M2-macrophage-reprogramming-via-CIRBP-modulation; (g) MM-CAR-T-resistance-via-M2-macrophage-stratification; (h) cross-cancer-CIRBP-validation-cohort (pan-cancer-CIRBP-expression + survival-correlation); (i) cold-shock-domain-CSD-drug-discovery-platform. The first deliverable should be a 3-page CIRBP-PKM2-RBP-target-nomination + multiple-myeloma-refractory-cancer + Warburg-vulnerability + M2-macrophage-reprogramming brief, not a clinical-recommendation. Cross-validate with the BB AI-DD + cancer-immunotherapy + longevity-inflammaging service lanes for cross-service opportunity mapping. ### 10. Tracker note — Biotech Infrastructure / §20 silent-failure fix stable for 41+ consecutive days; HF intel probe absent today (no fabricated HF metrics); score-9 ladder fully-recycled again, score-7 + score-6 band delivers 3 fresh NOVEL signals across 3 BB lanes: The 2026-09-20 research-watcher scan returned 103 fresh hits across 27 queries (collected_at 2026-09-19T21:02Z, mtime 2026-09-20 06:02 KST, scan.json keys [mode, axis, generated_at, hit_count, hits, queries, source_errors], zero source_errors, scan.log with `[brown-biotech] watcher done` footer) after the §20 silent-failure fix has now been stable for 41+ consecutive days (per the cron-silent-failure-diagnostics skill §20). Today is a Sunday — the §18 weekend-gap pattern will reduce Saturday/Sunday coverage to the smaller subset, Monday back to full 27-query weekday set. The score-9 top-band is FULLY RECYCLED at accession level — all 12 score-9 accessions appear in the prior id:97-id:101 retention window (regex-extracted 60 GSE accessions from page.tsx for robust multi-day dedupe per the §20-A pitfalls pattern validated 2026-08-27). The actionable NOVEL signals today come from the score-7 and score-6 bands after pdat-filter (≤30 days freshest primary, ≤90 days fallback) + prior-id:97-id:101-dedupe + mammalian-taxon-filter + BB-domain-keyword-filter (stromal-niche / EMT-state / TGFβ1 / Oncostatin-M / OSM / cancer-associated-fibroblast / spatial-transcriptomics / scRNA-seq / ATGL / PNPLA2 / fatty-acid-metabolism / plasma-membrane-cholesterol / phospholipid-remodeling / CIRBP / PKM2 / intron-retention / multiple-myeloma / macrophage-alternative-activation / M2-polarization / RNA-binding-protein / Warburg-effect). The 3 NOVEL signals today span 3 BB lanes: (a) **Bioinformatics & Multi-Omics + Cancer EMT + Two Distinct Stromal Niches + scRNA-seq + Spatial Companion Bundle** — GSE316348 + GSE316768 companion bundle (RNA-Seq + scRNA-seq, score 7/7, n=30 total, pdat 2026/09/17 = 2 days ago, selected as Signal 1), (b) **Longevity + Metabolic Disease + ATGL/PNPLA2 Target Nomination + Plasma-Membrane Cholesterol Homeostasis** — GSE346963 ATGL-dependent-phospholipid-remodeling-membrane-cholesterol atlas (selected as Signal 2, score 6, n=8, pdat 2026/09/18 = 1 day ago), (c) **AI Drug Discovery + Refractory Cancer + Multiple Myeloma + RNA-Binding-Protein Target Nomination + Macrophage Alternative Activation** — GSE339058 CIRBP-PKM2-intron-retention-MM-macrophage-reprogramming atlas (selected as Signal 3, score 6, n=4, pdat 2026/09/18 = 1 day ago). Honorable mentions: GSE347085 [Hepatitis E virus insertions antiviral response, n=12, pdat 2026/09/18], GSE347843 [rice OsPSY8 overexpression — out of BB-domain, plant], GSE316348+GSE316768 [companion bundle — selected as Signal 1], GSE286186 [Single-Nucleus Transcriptomic Atlas of Human Glomerular Diseases, n=59, pdat 2026/09/08], GSE337131 [Spatial transcriptomics of early-grade Huntington's disease Caudate, n=2, pdat 2026/09/01], GSE342738 [Multiscale spatial transcriptomics human amygdala Visium HD, n=5, pdat 2026/08/24], GSE314596 [FAP-Directed Immunotherapy atherosclerosis CITEseq, n=58, pdat 2026/02/02], GSE330687 [PD-L1 metabolic reprogramming NSCLC, n=6, pdat 2026/05/17]. HF intel layer continues to track 9 vendor-intel entries (longevity-db/human-muscle-aging-atlas-snRNAseq + mouse-muscle-aging-atlas-snRNAseq, ProteinGym_v1, OATML-Markslab/ProteinGym_v1, genbio-ai/ProteinGYM-DMS, ICML2022/ProteinGym, Jianwen/protein_ligand_cofolding_posebusters, SharkieJones/celldega_Visium-HD_hCRC). --- ## 💼 Next Steps - **[Open GSE316348 in GEO (TGFβ1+OSM stromal niches RNA-Seq, n=21)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE316348)** - **[Open GSE316768 in GEO (TGFβ1+OSM stromal niches scRNA-seq, n=9)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE316768)** - **[Open GSE346963 in GEO (ATGL phospholipid remodeling membrane cholesterol, n=8)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE346963)** - **[Open GSE339058 in GEO (CIRBP PKM2 intron retention multiple myeloma, n=4)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE339058)** - **[Request TGFβ1+OSM stromal niche EMT state control + spatial stromal cell state classifier brief](https://brownbio.tech/services/multiomics#brief)** - **[Request ATGL target nomination + membrane cholesterol homeostasis + fatty acid metabolism drug discovery brief](https://brownbio.tech/services/longevity#brief)** - **[Request CIRBP-PKM2 RBP target nomination + multiple myeloma + Warburg metabolism vulnerability brief](https://brownbio.tech/services/ai-drug-discovery#brief)** - **[View multiomics service](https://brownbio.tech/services/multiomics)** - **[View longevity service](https://brownbio.tech/services/longevity)** - **[View AI drug discovery service](https://brownbio.tech/services/ai-drug-discovery)** - **[View fibrosis service](https://brownbio.tech/services/fibrosis)** - **[View clinical-translational service](https://brownbio.tech/services/ai-drug-discovery)** --- ## 📡 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:** 2026-09-20 06:07 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 — 2026-09-20
PubMed/GEO scan · research-watcher · 06:00 KST