Observed Antibody Space (OAS) paired heavy and light chains, 2023-09-09 snapshot
- Provider
- Oxford Protein Informatics Group (OAS); snapshot by Brian Loyal
- License
- CC-BY-4.0 (source)
- Size
- 207 files
- Format
- Parquet (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/bloyal/oas-paired-sequence-data)
- Access
- Free and public. No account with the data provider is needed.
- Category
- Biology
- Source
- https://opig.stats.ox.ac.uk/webapps/oas/oas_paired/
About Observed Antibody Space (OAS) paired heavy and light chains, 2023-09-09 snapshot
Paired heavy- and light-chain antibody sequences from the paired section of the Observed Antibody Space (OAS), downloaded from the Oxford Protein Informatics Group on 9 September 2023. Paired data comes from single-cell repertoire sequencing, so each record keeps a heavy chain with the light chain from the same B cell, which is what antibody structure and developability models need and what bulk repertoires lack. The repo is organised by species: human (about 1.69 million pairs), mouse BALB/c, mouse C57BL/6 and rat SD, with the cleaning script used to build it. OAS data is CC BY 4.0; cite Olsen, Boyles and Deane, Protein Science 2022.
How to load this dataset in Python
from datasets import load_dataset
ds = load_dataset("bloyal/oas-paired-sequence-data")
print(ds)Data structure
File types in sampled files: Parquet
Files
Example file: human/train_0.parquet
Analyze this dataset without downloading it
On Scigantic, Observed Antibody Space (OAS) paired heavy and light chains, 2023-09-09 snapshot opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Biology datasets
- DynaCLR: live-cell infection dynamics and single-cell tracks
- ChronoRoot 2.0 segmentation dataset: annotated infrared images of Arabidopsis and tomato seedlings
- ESM-2 650M protein language model weights (esm2_t33_650M_UR50D)
- An MGRS-based 4km grid for the FTW Benchmark Dataset
- PRISM Antibody Training Data: 66 Million Unpaired and 764k Paired Human Antibody Sequences from OAS with Germline Labels (Romero Lab, Duke)
- PDB seqres FASTA 2022-09-28 (AlphaFold 3 training set)
- RNAcentral clustered ncRNA 90% identity (AlphaFold 3 training set)
- PDB mmCIF structures 2022-09-28 (AlphaFold 3 training set)