Update README.md
Browse files
README.md
CHANGED
@@ -48,8 +48,8 @@ license: cc-by-nc-nd-4.0
|
|
48 |
# Dataset Card for The Cancer Genome Atlas (TCGA) Multimodal Dataset
|
49 |
<!-- Provide a quick summary of the dataset. -->
|
50 |
|
51 |
-
The Cancer Genome Atlas (TCGA) Multimodal Dataset is a comprehensive collection of clinical data, pathology reports, and
|
52 |
-
This dataset aims to facilitate research in multimodal machine learning for oncology by providing embeddings generated using state-of-the-art models
|
53 |
|
54 |
- **Curated by:** Lab Rasool
|
55 |
- **Language(s) (NLP):** English
|
@@ -61,13 +61,27 @@ This dataset aims to facilitate research in multimodal machine learning for onco
|
|
61 |
```python
|
62 |
from datasets import load_dataset
|
63 |
|
64 |
-
|
65 |
-
|
66 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
wsi_dataset = load_dataset("Lab-Rasool/TCGA", "wsi", split="uni")
|
|
|
|
|
68 |
molecular_dataset = load_dataset("Lab-Rasool/TCGA", "molecular", split="senmo")
|
69 |
-
|
70 |
-
|
|
|
|
|
71 |
```
|
72 |
|
73 |
Example code for loading HF dataset into a PyTorch Dataloader.
|
@@ -95,20 +109,49 @@ class CustomDataset(Dataset):
|
|
95 |
|
96 |
if __name__ == "__main__":
|
97 |
|
98 |
-
|
|
|
|
|
99 |
wsi_dataset = load_dataset("Lab-Rasool/TCGA", "wsi", split="uni")
|
100 |
|
101 |
-
|
102 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
103 |
break
|
104 |
```
|
105 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
106 |
## Dataset Creation
|
107 |
|
108 |
#### Data Collection and Processing
|
109 |
The raw data for this dataset was acquired using MINDS, a multimodal data aggregation tool developed by Lab Rasool.
|
110 |
The collected data includes clinical information, pathology reports, and whole slide images from The Cancer Genome Atlas (TCGA).
|
111 |
-
The embeddings were generated using the HoneyBee embedding processing tool, which utilizes foundational models such as GatorTron and UNI.
|
112 |
|
113 |
#### Who are the source data producers?
|
114 |
The source data for this dataset was originally collected and maintained by The Cancer Genome Atlas (TCGA) program, a landmark cancer genomics project jointly managed by the National Cancer Institute (NCI).
|
|
|
48 |
# Dataset Card for The Cancer Genome Atlas (TCGA) Multimodal Dataset
|
49 |
<!-- Provide a quick summary of the dataset. -->
|
50 |
|
51 |
+
The Cancer Genome Atlas (TCGA) Multimodal Dataset is a comprehensive collection of clinical data, pathology reports, slide images, molecular data, and radiology images for cancer patients.
|
52 |
+
This dataset aims to facilitate research in multimodal machine learning for oncology by providing embeddings generated using state-of-the-art models including GatorTron, MedGemma, Qwen, Llama, UNI, SeNMo, REMEDIS, and RadImageNet.
|
53 |
|
54 |
- **Curated by:** Lab Rasool
|
55 |
- **Language(s) (NLP):** English
|
|
|
61 |
```python
|
62 |
from datasets import load_dataset
|
63 |
|
64 |
+
# Clinical data embeddings (4 models available)
|
65 |
+
clinical_gatortron = load_dataset("Lab-Rasool/TCGA", "clinical", split="gatortron")
|
66 |
+
clinical_medgemma = load_dataset("Lab-Rasool/TCGA", "clinical", split="medgemma")
|
67 |
+
clinical_qwen = load_dataset("Lab-Rasool/TCGA", "clinical", split="qwen")
|
68 |
+
clinical_llama = load_dataset("Lab-Rasool/TCGA", "clinical", split="llama")
|
69 |
+
|
70 |
+
# Pathology report embeddings (4 models available)
|
71 |
+
pathology_gatortron = load_dataset("Lab-Rasool/TCGA", "pathology_report", split="gatortron")
|
72 |
+
pathology_medgemma = load_dataset("Lab-Rasool/TCGA", "pathology_report", split="medgemma")
|
73 |
+
pathology_qwen = load_dataset("Lab-Rasool/TCGA", "pathology_report", split="qwen")
|
74 |
+
pathology_llama = load_dataset("Lab-Rasool/TCGA", "pathology_report", split="llama")
|
75 |
+
|
76 |
+
# Whole slide image embeddings
|
77 |
wsi_dataset = load_dataset("Lab-Rasool/TCGA", "wsi", split="uni")
|
78 |
+
|
79 |
+
# Molecular data embeddings
|
80 |
molecular_dataset = load_dataset("Lab-Rasool/TCGA", "molecular", split="senmo")
|
81 |
+
|
82 |
+
# Radiology embeddings (2 models available)
|
83 |
+
radiology_remedis = load_dataset("Lab-Rasool/TCGA", "radiology", split="remedis")
|
84 |
+
radiology_radimagenet = load_dataset("Lab-Rasool/TCGA", "radiology", split="radimagenet")
|
85 |
```
|
86 |
|
87 |
Example code for loading HF dataset into a PyTorch Dataloader.
|
|
|
109 |
|
110 |
if __name__ == "__main__":
|
111 |
|
112 |
+
# Load clinical embeddings from different models
|
113 |
+
clinical_gatortron = load_dataset("Lab-Rasool/TCGA", "clinical", split="gatortron")
|
114 |
+
clinical_llama = load_dataset("Lab-Rasool/TCGA", "clinical", split="llama")
|
115 |
wsi_dataset = load_dataset("Lab-Rasool/TCGA", "wsi", split="uni")
|
116 |
|
117 |
+
# Example: Access embeddings
|
118 |
+
for index, item in enumerate(clinical_gatortron):
|
119 |
+
embedding = np.frombuffer(item.get("embedding"), dtype=np.float32).reshape(item.get("embedding_shape"))
|
120 |
+
print(f"GatorTron embedding shape: {embedding.shape}") # Shape: (1024,)
|
121 |
+
break
|
122 |
+
|
123 |
+
for index, item in enumerate(clinical_llama):
|
124 |
+
embedding = np.frombuffer(item.get("embedding"), dtype=np.float32).reshape(item.get("embedding_shape"))
|
125 |
+
print(f"Llama embedding shape: {embedding.shape}") # Shape: (2304,)
|
126 |
break
|
127 |
```
|
128 |
|
129 |
+
## Dataset Statistics
|
130 |
+
|
131 |
+
### Clinical Data
|
132 |
+
- **10,771 patient records** per model
|
133 |
+
- **113 columns** including clinical metadata and embeddings
|
134 |
+
- **Embedding dimensions:**
|
135 |
+
- GatorTron: 1024
|
136 |
+
- MedGemma: 2560
|
137 |
+
- Qwen: 1024
|
138 |
+
- Llama: 2304
|
139 |
+
|
140 |
+
### Pathology Reports
|
141 |
+
- **10,857 patient records** per model
|
142 |
+
- **17 columns** including pathology metadata and embeddings
|
143 |
+
- **Embedding dimensions:**
|
144 |
+
- GatorTron: 1024
|
145 |
+
- MedGemma: 2560
|
146 |
+
- Qwen: 1024
|
147 |
+
- Llama: 2304
|
148 |
+
|
149 |
## Dataset Creation
|
150 |
|
151 |
#### Data Collection and Processing
|
152 |
The raw data for this dataset was acquired using MINDS, a multimodal data aggregation tool developed by Lab Rasool.
|
153 |
The collected data includes clinical information, pathology reports, and whole slide images from The Cancer Genome Atlas (TCGA).
|
154 |
+
The embeddings were generated using the HoneyBee embedding processing tool, which utilizes foundational models such as GatorTron, MedGemma, Qwen, Llama, and UNI.
|
155 |
|
156 |
#### Who are the source data producers?
|
157 |
The source data for this dataset was originally collected and maintained by The Cancer Genome Atlas (TCGA) program, a landmark cancer genomics project jointly managed by the National Cancer Institute (NCI).
|