ML-Guided SARS-CoV-2 VHH Nanobody Design: Yeast-Display Screens, BLI Affinities and Neutralisation Curves
- Provider
- Google Research with Lumen Bioscience and Adaptive Biotechnologies (Angermueller, Mariet, Colwell et al.)
- License
- CC-BY-4.0 (source)
- Stored at
- Google Cloud Storage, gs://gresearch/covid_vhh_design (us)
- Access
- Free and public. No account with the data provider is needed.
- DOI
- 10.1101/2023.12.01.569227
- Category
- Drug Discovery
- Source
- https://doi.org/10.1101/2023.12.01.569227
About this dataset
Experimental measurements behind the machine-learning-guided design of single-domain antibodies (VHHs, nanobodies) against SARS-CoV-1 and SARS-CoV-2 by Angermueller, Mariet and colleagues. Three successive design rounds are provided as round1.csv.gz, round2.csv.gz and round3.csv.gz: each row is a VHH sequence with its parent, number of mutations from the parent (4 to 15 for the designed variants), design group (singles, MBO, LightGBM/mbo_lgb and other model-proposed sets), replicate and the measured binding score against a panel of antigen targets including SARS-CoV-1 RBD, SARS-CoV-2 Wuhan RBD and spike, and the Delta and Omicron BA.1 variants, alongside positive and negative controls. Orthogonal biolayer interferometry results are in bli_v1.csv, bli_v2.csv and bli_v2_curves.csv.gz with IC50, KD, ka and kdis per sequence-target pair, plus bli_v2_neutralization_curves.csv.gz for pseudovirus neutralisation. parent_seq.csv gives the starting VHH, aaindex_components.csv the amino-acid featurisation, and lgb_model/ and model_scores/ the trained LightGBM scorer and its predictions. Measured 2026-11-15: 42 objects, about 50 MB, CSV and gzipped CSV. Load with pandas or the covid_vhh_design package. All datasets in the google-research repository are released CC BY 4.0; cite Angermueller et al. 2023.
How to access this dataset in Python
import gcsfs
fs = gcsfs.GCSFileSystem(token='anon')
fs.ls('gresearch/covid_vhh_design/', detail=False)[:10]Lists the files with anonymous access. Runs on any machine, with no cloud account.
Data structure
Files
Example file: covid_vhh_design/lgb_model/classifier/0000.txt
Analyze this dataset without downloading it
On Scigantic, ML-Guided SARS-CoV-2 VHH Nanobody Design: Yeast-Display Screens, BLI Affinities and Neutralisation Curves 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 Drug Discovery datasets
- Morphing libraries, QSAR models, and compounds predicted to be active on the Glucocorticoid receptor (GR)
- Metabolomics WorkBench Compound Dictionary
- FragBERTa: A Fragment-based Molecular Representation Learning Model with Sequential Attachment-based Fragment Embeddings - Models and Datasets
- Dataset for evaluating accuracy of Boltz-2's structure and binding affinity predictions
- Dataset for: Image-based screen capturing misfolding status of Niemann-Pick type C1 identifies potential candidates for chaperone drugs
- Novartis CYP3A4 Time-Dependent Inhibition (TDI) ADMET dataset
- Consensus machine-learning models for protein-ligand binding affinity estimation
- Shared molecular regulation of quiescence in neural and glioma stem cells reveals therapeutic vulnerabilities