Modeling brain tumor cell states with AI virtual cells.用 AI 虚拟细胞刻画脑肿瘤细胞状态。

GlioState is a curated, source-verified knowledge hub on AI virtual cells, read through the lens of neuro-oncology: glioblastoma and glioma, other brain tumors, brain metastases — and the largely unmapped niche of spinal cord tumors.GlioState 是一个经过逐条核验的 AI 虚拟细胞知识库，以神经肿瘤学视角整理：胶质母细胞瘤与胶质瘤、其他脑肿瘤、脑转移瘤，以及几乎空白的脊髓肿瘤方向。

Browse the library (139)浏览文献库（139 条）Brain tumor view脑肿瘤专题Review: GBM cell states综述：GBM 细胞状态

What is an AI virtual cell?什么是 AI 虚拟细胞？

As defined by Bunne et al. (Cell, 2024), an AI virtual cell is built on multimodal, multiscale, universal representations of cells; it aims to simulate in silico how cell states change under arbitrary perturbations and to act as a “virtual instrument” that guides real experiments.Bunne 等（Cell，2024）提出：AI 虚拟细胞以多模态、多尺度的通用细胞表征为基础，旨在计算机中模拟细胞在任意扰动下的状态变化，并作为“虚拟实验仪器”指导真实实验。 VC0001

Benchmarks also show that current deep models often fail to beat simple linear baselines on perturbation prediction — so we track critiques as carefully as claims.同时，多项基准研究显示当前深度模型在扰动预测上常常不如简单线性基线——因此我们同样认真追踪批评与反证。 VC0048

Our mission我们的使命

  • Every item verified against PubMed, Europe PMC, Crossref or an official page.每条条目均对照 PubMed、Europe PMC、Crossref 或官方页面核验。
  • A clinician-researcher’s view of what matters for brain and spinal cord tumors.以临床医生-研究者视角，聚焦脑与脊髓肿瘤真正重要的问题。
  • Bilingual (English / 中文), updated weekly, open to read.中英双语、每周更新、免费开放。
  • Not medical advice; not a recommendation.不构成医疗建议，也不构成推荐。

47 items are directly about brain tumors, 18 partly relevant.其中 47 条直接涉及脑肿瘤，18 条部分相关。

Research network研究知识网络

All research hubs →全部研究专题 →

Brain tumor highlights脑肿瘤重点条目

All brain tumor items →全部脑肿瘤条目 →

Latest additions最新条目

Full library →完整文献库 →
preprint预印本✓ Verified已核验VC0139

Engineered regulatory states resolve context-dependent genetics in macular degeneration

Metzner E, Guruvayurappan K, Southard KM, Ardy RC, Dey KK, Norman TM. · bioRxiv · 2026-10-12

Norman/Dey labs (MSKCC): CRISPRa Multiome Perturb-seq of 119 TFs + dual-guide Perturb-seq of 1,225 TF pairs in RPE-1; chromatin priming predicts emergent states; Mosaic (Transformer) predicts TF-pair interaction residuals (r=0.484) and nominated a validated triple-TF combination.Norman/Dey 实验室（MSKCC）：在 RPE-1 中对 119 个 TF 做 CRISPRa Multiome Perturb-seq、对 1,225 个 TF 对做双向导 Perturb-seq；染色质预激活可预测新兴状态；Mosaic（Transformer）预测 TF 对相互作用残差（r=0.484），并提名经实验验证的三 TF 组合。

DOI

Upcoming conferences近期会议

Calendar →会议日历 →

Frequently asked questions常见问题

What is an AI virtual cell?什么是 AI 虚拟细胞？

An AI virtual cell is a computational model intended to simulate how a cell’s state changes under perturbations such as gene edits or drugs. Bunne et al. (Cell, 2024) define it as built on multimodal, multiscale, universal cell representations that can act as a “virtual instrument” to guide experiments. A “world model” framing of the virtual cell has since been proposed. It remains a research goal rather than a finished, validated tool.AI 虚拟细胞是一种计算模型，旨在模拟细胞在基因编辑、药物等扰动下的状态变化。Bunne 等（Cell，2024）将其定义为基于多模态、多尺度通用细胞表征、可作为“虚拟实验仪器”指导实验的模型；此后又有人提出虚拟细胞的“世界模型”框架。它目前仍是研究目标，而非已完成验证的工具。 VC0001 VC0131

Do current virtual cell models actually work?现有虚拟细胞模型真的有效吗？

Only partly, according to current benchmarks. Several studies report that deep-learning perturbation models do not yet reliably outperform simple linear baselines, and zero-shot performance of single-cell foundation models is limited. The Virtual Cell Challenge was launched as a “Turing test” to measure progress on exactly these questions.依据现有基准研究，答案是“部分有效”。多项研究显示，深度学习扰动预测模型尚不能稳定超越简单线性基线，单细胞基础模型的零样本表现也有限。“虚拟细胞挑战赛”正是作为一种“图灵测试”来衡量这方面的进展。 VC0048 VC0049 VC0050 VC0002

What is GlioState?GlioState 是什么？

GlioState is a bilingual (English/Chinese) knowledge hub that curates 139 source-verified items on AI virtual cells, read from a neuro-oncology perspective. It is edited by the GlioState Editorial Team from a neuro-oncology perspective. It is for research and education only and is not medical advice.GlioState 是一个中英双语知识库，从神经肿瘤学视角收录 139 条经逐条核验的 AI 虚拟细胞相关条目，由 GlioState 编辑团队主编。仅供科研与教育参考，不构成医疗建议。