Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language: en
|
| 3 |
+
license: mit
|
| 4 |
+
task_categories:
|
| 5 |
+
- tabular-classification
|
| 6 |
+
- feature-extraction
|
| 7 |
+
tags:
|
| 8 |
+
- bioinformatics
|
| 9 |
+
- cancer-research
|
| 10 |
+
- glioblastoma
|
| 11 |
+
- survival-prediction
|
| 12 |
+
- gene-expression
|
| 13 |
+
pretty_name: GBM Survival Gene Expression Dataset
|
| 14 |
+
size_categories:
|
| 15 |
+
- 1GB<n<10GB
|
| 16 |
+
---
|
| 17 |
+
|
| 18 |
+
# 🧬 GBM Survival Gene Expression Dataset
|
| 19 |
+
|
| 20 |
+
This dataset contains large-scale gene expression data (RNA-Seq and Microarray) curated for predicting the survival outcomes of patients with Glioblastoma Multiforme (GBM) and Lower-Grade Glioma (LGG).
|
| 21 |
+
|
| 22 |
+
It is specifically designed to be used with the [GBM Survival Mamba](https://github.com/congkx123789/Tin_sinh_h-c) project.
|
| 23 |
+
|
| 24 |
+
## 📂 Data Structure
|
| 25 |
+
|
| 26 |
+
The data is organized into three main phases:
|
| 27 |
+
|
| 28 |
+
### 1. `01_raw/` (Raw Data)
|
| 29 |
+
Untouched genomic data from international portals:
|
| 30 |
+
- **TCGA**: RNA-Seq data from The Cancer Genome Atlas (USA).
|
| 31 |
+
- **CGGA**: Chinese Glioma Genome Atlas data (Asia).
|
| 32 |
+
- **GEO**: Various cohorts from NCBI Gene Expression Omnibus (REMBRANDT, GSE4412, etc.).
|
| 33 |
+
|
| 34 |
+
### 2. `02_processed/` (Cleaned & Normalized)
|
| 35 |
+
Ready-to-use data for machine learning:
|
| 36 |
+
- **training/**: Internal Train/Val/Test splits used for model development.
|
| 37 |
+
- **validation/**: External independent cohorts for performance verification.
|
| 38 |
+
- Files include `X.csv` (feature matrices with 16,383 genes) and `y.csv` (survival time and status).
|
| 39 |
+
|
| 40 |
+
### 3. `03_metadata/`
|
| 41 |
+
Mapping tables between Gene Symbols and Probe IDs, as well as clinical metadata.
|
| 42 |
+
|
| 43 |
+
## 🚀 How to use with the Project
|
| 44 |
+
|
| 45 |
+
If you have cloned the [GitHub repository](https://github.com/congkx123789/Tin_sinh_h-c), you can automatically sync this data by running:
|
| 46 |
+
|
| 47 |
+
```bash
|
| 48 |
+
python scripts/download_data_hf.py
|
| 49 |
+
```
|
| 50 |
+
|
| 51 |
+
This will download all files and place them in the correct `data/` directory structure.
|
| 52 |
+
|
| 53 |
+
## 📝 Data Sources & Citations
|
| 54 |
+
|
| 55 |
+
If you use this data, please cite the original sources:
|
| 56 |
+
1. TCGA Research Network: https://www.cancer.gov/tcga
|
| 57 |
+
2. CGGA Database: http://www.cgga.org.cn/
|
| 58 |
+
3. NCBI GEO: https://www.ncbi.nlm.nih.gov/geo/
|
| 59 |
+
|
| 60 |
+
**Project Maintainer**: Cong, K. X.
|