Duplicate from InstaDeepAI/CoVUniBind

Co-authored-by: Francesco Saccon <fsaccon@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,61 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text
+sars_binding_dataset_paper_data_dms_bloom_ab_escape.csv filter=lfs diff=lfs merge=lfs -text
+sars_binding_dataset_paper_data_dms_cao_ab_escape.csv filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,118 @@
+---
+tags:
+- antibody
+- antigen
+- binding
+- COVID-19
+- SARS-CoV-2
+- benchmark
+pretty_name: CoV-UniBind
+---
+<div align="center">
+
+<h1>CoV-UniBind - Coronavirus Unified Binding Database</h1>
+
+</div>
+
+<p align="center">
+  <img src="https://cdn-uploads.huggingface.co/production/uploads/681b661257cfafb5ead152d1/WvLoDuemHIwY8bQfOMahb.png" alt="Description" width="300"/>
+</p>
+
+**CoV-UniBind** curates and integrates structural and biochemical data on antibodies specifically elicited by SARS-CoV-2 and
+related coronaviruses. It links 3D antibody structures to their binding properties across viral variants, incorporating
+epitope and sequence information. This dataset serves as a comprehensive resource for analysing antibody-antigen
+interactions in the context of COVID-19 and provides a foundation for binding classification and regression tasks using
+deep learning models.
+
+## Database Overview
+
+| Dataset | Description | Label | Type | References (doi) |
+|------|-------------|-------|-----| ------- |
+| covabdab | Binding labels from the Coronavirus Antibody Database (CoV-AbDab) | binding | bool | 10.1093/bioinformatics/btaa739 |
+| dms_bloom | Deep Mutational Scanning escape data from Greaney et al. 2022 | escape | float | 10.1038/s41564-021-00972-2; 10.1016/j.chom.2020.11.007; 10.1038/s41467-021-24435-8; 10.1016/j.xcrm.2021.100255; 10.1126/science.abf9302; 10.1038/s41467-021-24435-8; 10.1038/s41564-021-00972-2; 10.1038/s41586-021-04385-3; 10.1038/s41586-022-04980-y; 10.1038/s41586-021-03817-4; 10.1038/s41586-021-03807-6 |
+| dms_cao | Deep Mutational Scanning escape data from Cao et al. 2023 | escape | float | 10.1038/s41586-022-05644-7 |
+| jian_elisa | ELISA antibody IC50 measurements from Jian et al. 2025 | IC50 | float | 10.1038/s41586-024-08315-x |
+| spr | Surface Plasmon Resonance antibody affinity measurements from multiple sources | KD | float | 10.1038/s41467-024-54916-5; 10.1038/s41586-020-2380-z; 10.1038/s41467-021-24514-w; 10.1016/j.immuni.2022.06.005; 10.1016/j.xcrm.2023.100991 |
+| drdb | Neutralisation potency data from the SARS-CoV-2 Resistance Database (DRDB) | IC50 | float | 10.1371/journal.pone.0261045 |
+
+## Database Structure
+```
+.
+├── antibody_info
+│   └── antibody_synonyms.csv
+├── data
+│   ├── covabdab_binding.parquet
+│   │   └── structures
+│   │       ├── processed.zip
+│   │       └── trimmed.zip
+│   ├── dms_bloom_ab_escape.parquet
+│   ├── dms_cao_ab_escape.parquet
+│   ├── drdb_binding_potency.parquet
+│   ├── jian_elisa_ab_ic50.parquet
+│   └── spr_ab_affinity.parquet
+├── scores
+│   ├── covabdab_binding_scores.parquet
+│   ├── dms_bloom_ab_escape_scores.parquet
+│   ├── dms_cao_ab_escape_scores.parquet
+│   ├── drdb_binding_potency_scores.parquet
+│   ├── jian_elisa_ab_ic50_scores.parquet
+│   └── spr_ab_affinity_scores.parquet
+├── cov-unibind.py
+└── README.md
+```
+
+## Usage Guide
+
+```python
+from datasets import load_dataset
+
+data='drbd' # specify the dataset name based on table above
+
+dataset = load_dataset("InstaDeepAI/cov-unibind",name=data)
+```
+
+
+## Dataset Schema
+
+The table below includes information about the columns contained in the datasets.
+
+| Column Name | Description | Type | Nullable | Example |
+|---|---|---|---|---|
+| `antibody_name` | Name of the antibody | *str* | False | `bd30_515;bd_515` |
+| `antigen_lineage` | Antigen lineage | *str* | False | `BA.1` |
+| `target_value` | Experimental binding value| *float* or *bool* | False | `-2.327902` or `True` |
+| `target_type` | Type of target value | *str* | False | `IC50_log10_fold` |
+| `source_name` | Source of the data | *str* | False | `jian_2024_nature` |
+| `source_doi` | DOI of the source | *str* | False | `10.1038/s41586-024-08315-x` |
+| `assay_name` | Name of the assay | *str* | False | `elisa` |
+| `pdb_id` | PDB structure ID | *str* | False | `7e88` |
+| `structure_release_date` | Release date of the structure | *str* | False | `03/01/21` |
+| `structure_resolution` | Resolution of the structure (Å) | *float* | False | `3.14` |
+| `mutations` | Lineage consensus mutations | *str* | False | `A67V H69- V70- T95I...` |
+| `antigen_chain_ids` | Chain IDs of the antigen | *str* | False | `C` |
+| `antigen_domain` | Domain of the antigen | *str* | False | `RBD` |
+| `antigen_residue_indices` | Residue indices of the antigen | *str* | False | `(13, 568)` |
+| `antigen_residue_indices_trimmed` | Antigen residue indices, trimmed | *float* | True | `(333, 526)` |
+| `antigen_host` | Host organism of the antigen | *str* | False | `severe acute respiratory syndrome coronavirus 2 (2697049)` |
+| `antibody_heavy_chain_id` | Heavy chain ID of the antibody | *str* | False | `C` |
+| `antibody_light_chain_id` | Light chain ID of the antibody | *str* | False | `B` |
+| `epitope_residues` | Residues of the epitope | *str* | False | `R403 D405 T415` |
+| `epitope_mutations` | PDB antigen mutations in the epitope | *str* | True | `D405T` |
+| `epitope_domain` | Spike domain where the antibody binds | *str* | False | `RBD` |
+| `epitope_alteration_count` | Number of alterations in the epitope | *float* | True | `2` |
+| `spike_sequence` | Full spike protein sequence | *str* | False | `MFVFLVLLPLVSSQCVNLTTRTQL...` |
+| `antibody_heavy_chain_sequence` | Sequence of the antibody heavy chain | *str* | False | `EVQLVESGGGLVQPGGSLRLSCAASEFIVSRNYMSWVR...` |
+| `antibody_light_chain_sequence` | Sequence of the antibody light chain | *str* | False | `DIQMTQSPSSLSASVGDRVTITCQASQDINKYLNWYQQK...` |
+| `antibody_vh_sequence` | Sequence of the VH domain | *str* | False | `EVQLVESGGGLVQPGGSLRLSCAASEFIVSRNYMSWVRQ...` |
+| `antibody_vl_sequence` | Sequence of the VL domain | *str* | False | `DIQMTQSPSSLSASVGDRVTITCQASQDINKYLNW...` |
+| `antigen_sequence` | Full antigen sequence | *str* | False | `TNLCPFDEVF-NATRFASVYAWNR--KRISNCVADYSVLYNLAPFFTFKCYGVSP...` |
+| `antigen_sequence_trimmed` | Trimmed antigen sequence | *float* | True | `TRFASV-YAWNRKRISNCVADYSVLYNLAPFFT-FKCYGVSP...` |
+| `antigen_sequence_without_indels`| Antigen sequence without insertions/deletions | *str* | False | `TNLCPFDEVFNATRFASVYAWNRKRISNCVAD...` |
+| `antigen_sequence_trimmed_without_indels` | Trimmed antigen sequence without insertions/deletions | *float* | True | `NATRFASVYAWNRKRISNCVAD...` |
+| `antigen_pdb_sequence` | Antigen sequence from PDB | *str* | False | `TNLCPFGEVFNATRFASVYAWNRKRISNCVADYSVLYNSA...` |
+| `antigen_pdb_sequence_trimmed` | Trimmed antigen sequence from PDB | *float* | True | `NATRFASVYAWNRKRISNCVADYSVLYNSA...` |
+
+## Acknowledgments
+
+This project makes use of publicly available antibody datasets listed above. We acknowledge the contributions by the teams
+responsible for compiling and maintaining these valuable resources.

antibody_info/antibody_synonyms.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1c440450dfd58a34adcafb08d43fa80d9a5c72550ce46dec8d41998673f571a
+size 23437

cov-unibind.py

@@ -0,0 +1,131 @@
+import pandas as pd
+import datasets
+
+_DESCRIPTION = """\
+Multi-source dataset of antibody-mutation interactions including IC50, binding, escape, and affinity measurements.
+Also includes antibody synonyms with CDR sequences and epitope information.
+"""
+
+_FEATURES = {
+    'antibody_name': datasets.Value("string"),
+    'antigen_lineage': datasets.Value("string"),
+    'target_value': datasets.Value("float"),
+    'target_type': datasets.Value("string"),
+    'source_name': datasets.Value("string"),
+    'source_doi': datasets.Value("string"),
+    'assay_name': datasets.Value("string"),
+    'pdb_id': datasets.Value("string"),
+    'structure_release_date': datasets.Value("string"),
+    'structure_resolution': datasets.Value("float"),
+    'mutations': datasets.Value("string"),
+    'antigen_chain_ids': datasets.Value("string"),
+    'antigen_domain': datasets.Value("string"),
+    'antigen_residue_indices': datasets.Value("string"),
+    'antigen_residue_indices_trimmed': datasets.Value("string"),
+    'antigen_host': datasets.Value("string"),
+    'antibody_heavy_chain_id': datasets.Value("string"),
+    'antibody_light_chain_id': datasets.Value("string"),
+    'epitope_residues': datasets.Value("string"),
+    'epitope_mutations': datasets.Value("string"),
+    'epitope_domain': datasets.Value("string"),
+    'epitope_alteration_count': datasets.Value("string"),
+    'spike_sequence': datasets.Value("string"),
+    'antibody_heavy_chain_sequence': datasets.Value("string"),
+    'antibody_light_chain_sequence': datasets.Value("string"),
+    'antibody_vh_sequence': datasets.Value("string"),
+    'antibody_vl_sequence': datasets.Value("string"),
+    'antigen_sequence': datasets.Value("string"),
+    'antigen_sequence_trimmed': datasets.Value("string"),
+    'antigen_sequence_without_indels': datasets.Value("string"),
+    'antigen_sequence_trimmed_without_indels': datasets.Value("string"),
+    'antigen_pdb_sequence': datasets.Value("string"),
+    'antigen_pdb_sequence_trimmed': datasets.Value("string"),
+}
+
+_ANTIBODY_SYNONYMS_FEATURES = {
+    'antibody_name': datasets.Value("string"),
+    'pdb_id': datasets.Value("string"),
+    'antibody_heavy_chain_cdr1': datasets.Value("string"),
+    'antibody_heavy_chain_cdr2': datasets.Value("string"),
+    'antibody_heavy_chain_cdr3': datasets.Value("string"),
+    'antibody_light_chain_cdr1': datasets.Value("string"),
+    'antibody_light_chain_cdr2': datasets.Value("string"),
+    'antibody_light_chain_cdr3': datasets.Value("string"),
+    'epitope_residues': datasets.Value("string"),
+    'epitope_domain': datasets.Value("string"),
+}
+
+_TABLES = {
+    "antibody_synonyms": {
+        "file": "antibody_info/antibody_synonyms_with_epitopes.parquet",
+        "features": _ANTIBODY_SYNONYMS_FEATURES,
+    },
+    "drdb": {
+        "file": "data/drdb_binding_potency.parquet",
+        "features": {
+            **_FEATURES,
+        }
+    },
+    "covabdab": {
+        "file": "data/covabdab_binding.parquet",
+        "features": {
+            **{
+                **_FEATURES,
+                "target_value": datasets.Value("bool"),
+            }
+        }
+    },
+    "dms_bloom": {
+        "file": "data/dms_bloom_ab_escape.parquet",
+        "features": {
+            **_FEATURES,
+        }
+    },
+    "dms_cao": {
+        "file": "data/dms_cao_ab_escape.parquet",
+        "features": {
+            **_FEATURES,
+        }
+    },
+    "jian_elisa": {
+        "file": "data/jian_elisa_ab_ic50.parquet",
+        "features": {
+            **_FEATURES,
+        }
+    },
+    "spr": {
+        "file": "data/spr_ab_affinity.parquet",
+        "features": {
+            **_FEATURES,
+        }
+    }
+}
+
+class CovUniBindConfig(datasets.BuilderConfig):
+    def __init__(self, **kwargs):
+        super().__init__(version=datasets.Version("1.0.0"), **kwargs)
+
+
+class CovUniBind(datasets.GeneratorBasedBuilder):
+    BUILDER_CONFIGS = [
+        CovUniBindConfig(name=table, description=f"{table} subset") for table in _TABLES
+    ]
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(_TABLES[self.config.name]["features"]),
+        )
+
+    def _split_generators(self, dl_manager):
+        file_path = _TABLES[self.config.name]["file"]
+        data_path = dl_manager.download_and_extract(file_path)
+        return [
+            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": data_path}),
+        ]
+
+    def _generate_examples(self, filepath):
+        df = pd.read_parquet(filepath)
+        for idx, row in df.iterrows():
+            yield idx, row.to_dict()
+

data/covabdab_binding.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d223e22f34cc63405b860dba8b66ac2c2022bf49e1b93f2f9fa64d9fe32fe5e8
+size 120221

data/dms_bloom_ab_escape.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19ef15de05ab2bdf4a39cd0250327a95ebecf99264e397861f2da0306ad12b50
+size 10042875

data/dms_cao_ab_escape.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b42dbd8e255c57a22158efe28b081194a31a3863b0cd5d38cf805d79b3453d0
+size 1576603

data/drdb_binding_potency.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36599544eba6035359dfcd751680223cd2333f427667d5de0c53bf5037965e7d
+size 242949

data/jian_elisa_ab_ic50.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:261c03227315fe659049878905e843aa8621dcf5f341987f601b79d2f88decbd
+size 51072

data/spr_ab_affinity.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5fa44994c696356aeeab1c536c3fe3fbdf2a90dec106613468b639af1f5f72e
+size 119903

data/structures/processed.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c361bd05bc6e29bbc1d6c6754015fc0f8805e4c8ba077749ee926e8b018df00
+size 256717383

data/structures/trimmed.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64144e4ad11648a7db1812f1577c2db96deb96973d754402edddce145b57671e
+size 63004759

scores/covabdab_binding_scores.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5dda25d563caacd7e7ef7d9a3e84685973c0942c7d1faf7f0ea1c7113120eaa4
+size 226051

scores/dms_bloom_ab_escape_scores.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80f8bbf666630ccf8a758d53f26291a548e6dfcf9f349836c417cd32d57687a3
+size 2939536

scores/dms_cao_ab_escape_scores.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0cdf6d17912983b0d68a7225d1ad20dc34fb9bdd5259ef0e6fc6ba7dc2a10e0e
+size 1235985

scores/drdb_binding_potency_scores.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f91cc0bfe3b81ee292f99451f6b1c654a79a33aa2e0c0642b17865129877d169
+size 412775

scores/jian_elisa_ab_ic50_scores.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c51726bf57e9387bde2c1285628b6c51c8e4b702da1e2448c9557aadf8305b59
+size 79465

scores/spr_ab_affinity_scores.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20fa89eee2bbd7aa9880853f3fd92e3bf72b5242e957f28d7d2e939ac9c52388
+size 312135