Update README.md

README.md

@@ -1,67 +1,118 @@
----
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-  - split: test
-    path: data/test-*
-dataset_info:
-  features:
-  - name: Peptide
-    dtype: string
-  - name: Sequence
-    dtype: string
-  - name: HLA
-    dtype: string
-  - name: Length
-    dtype: int64
-  - name: Log_meas
-    dtype: float64
-  splits:
-  - name: train
-    num_bytes: 142687867
-    num_examples: 125536
-  - name: test
-    num_bytes: 35661056
-    num_examples: 31385
-  download_size: 7716040
-  dataset_size: 178348923
----
-IEDB dataset: http://tools.iedb.org/mhci/download/  binding_data_2013.zip
-
-https://raw.githubusercontent.com/s7776d/CapsNet-MHC/refs/heads/main/dataset/IEDB_dataset/training_set.txt
-
-Code for adaptation: https://github.com/vladimirkovacevic/nucl-protein/blob/main/notebooks/EDA_Datasets.ipynb
-
-
----
-dataset_info:
-  features:
-  - name: HLA
-    dtype: string
-  - name: peptide
-    dtype: string
-  - name: Log_meas
-    dtype: float64
-  - name: Length
-    dtype: int64
-  - name: Sequence
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 142687867
-    num_examples: 125536
-  - name: test
-    num_bytes: 35661056
-    num_examples: 31385
-  download_size: 7716027
-  dataset_size: 178348923
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-  - split: test
-    path: data/test-*
----
+---
+tags:
+- immunology
+- MHC
+- epitope
+- binding
+- peptide
+- HLA
+- biology
+- huggingscience
+- science
+pretty_name: IEDB MHC Class I Binding Dataset 2013
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: test
+    path: data/test-*
+dataset_info:
+  features:
+  - name: Peptide
+    dtype: string
+  - name: Sequence
+    dtype: string
+  - name: HLA
+    dtype: string
+  - name: Length
+    dtype: int64
+  - name: Log_meas
+    dtype: float64
+  splits:
+  - name: train
+    num_bytes: 142687867
+    num_examples: 125536
+  - name: test
+    num_bytes: 35661056
+    num_examples: 31385
+  download_size: 7716040
+  dataset_size: 178348923
+size_categories:
+- 1M<n<10M
+---
+
+# IEDB MHC Class I Binding Dataset 2013
+
+## Dataset Description
+
+### Dataset Summary
+
+The IEDB 2013 dataset contains experimentally measured binding affinities between peptides and Major Histocompatibility Complex (MHC) class I molecules. This dataset is derived from the Immune Epitope Database (IEDB), which is a freely available resource containing an extensive collection of experimentally measured immune epitopes and tools for epitope analysis and prediction.
+
+The dataset includes binding measurements for various Human Leukocyte Antigen (HLA) alleles and peptides of different lengths, making it valuable for training machine learning models to predict MHC-peptide binding affinities - a crucial task in immunoinformatics, vaccine design, and personalized medicine.
+
+### Supported Tasks
+
+- **Regression**: The dataset can be used to train models for predicting MHC class I binding affinities, which consists of predicting the binding strength between a peptide and an MHC molecule. Success on this task is typically measured by achieving low mean squared error (MSE) or high Pearson correlation coefficients.
+
+
+## Getting Started
+
+### Quick Start
+
+```python
+from datasets import load_dataset
+
+# Load the dataset
+dataset = load_dataset("vladak/iedb-2013")
+
+# View the first few examples
+print(dataset["train"][:5])
+
+# Get dataset info
+print(f"Dataset size: {len(dataset['train'])}")
+print(f"Features: {dataset['train'].features}")
+```
+
+## Dataset Structure
+
+### Data Instances
+
+Each instance in the dataset represents a single binding measurement experiment between a peptide and an HLA allele:
+
+```json
+{
+  "HLA": "HLA-A*02:01",
+  "Peptide": "SLYNTVATL",
+  "Log_meas": 2.45,
+  "Length": 9,
+  "Sequence": "SLYNTVATL"
+}
+```
+
+### Data Fields
+
+| Field Name | Type | Description |
+|------------|------|-------------|
+| `HLA` | `string` | HLA allele identifier using standard nomenclature (e.g., "HLA-A*02:01", "HLA-B*07:02") |
+| `Peptide` | `string` | Amino acid sequence of the peptide using single-letter codes |
+| `Log_meas` | `float64` | Logarithm of the measured binding affinity. Lower values indicate stronger binding |
+| `Length` | `int64` | Length of the peptide sequence in amino acids |
+| `Sequence` | `string` | Same as peptide field - amino acid sequence of the peptide |
+
+### Data Splits
+
+The dataset is divided in train split (126k rows) and test split (31.4k rows)-
+
+## Dataset Creation
+
+
+### Source Data
+
+The data originates from the Immune Epitope Database (IEDB), which systematically collects and curates immune epitope data from peer-reviewed scientific literature and direct data submissions from researchers. The specific data used in this dataset comes from the 2013 version of the IEDB binding data release.
+
+**Data Sources:**
+- Original IEDB binding data: http://tools.iedb.org/mhci/download/binding_data_2013.zip
+- Processed training set: https://raw.githubusercontent.com/s7776d/CapsNet-MHC/refs/heads/main/dataset/IEDB_dataset/training_set.txt
+- Processing code: https://github.com/vladimirkovacevic/nucl-protein/blob/main/notebooks/EDA_Datasets.ipynb