GDPa2: Antibody developability dataset of 18 VHH constructs

Contains developability assay data for 18 nanobodies (VHHs), each produced in VHH-His and VHH-Fc formats, measured with the PROPHET-Ab platform described for GDPa1. VHHs are of growing interest for intracellular targeting, multispecific formats, and CNS penetration.

Example usage

Using pandas:

import pandas as pd

# Login using e.g. `hf auth login` to access this dataset
df = pd.read_csv("hf://datasets/ginkgo-datapoints/GDPa2/GDPa2.1.csv")

Using Hugging Face datasets:

from datasets import load_dataset

# Login using e.g. `hf auth login` to access this dataset
ds = load_dataset("ginkgo-datapoints/GDPa2")

Data processing

The main table in this data (GDPa2.1.csv) has one row per construct (51 rows across VHH-Fc in HEK293F, VHH-Fc in ExpiCHO, and VHH-His in HEK293F). It joins the "Sequences" sheet of the full datasheet with the "Assay Data - average" sheet, which reports the mean, standard deviation, and replicate count for each assay.

Antibody production

VHH-Fc constructs were produced in HEK293F and ExpiCHO, and VHH-His constructs in HEK293F (see production_host and tag).

Developability assays

1. Hydrophobicity by HIC
2. Aggregation by SEC (% monomer)
3. Colloidal stability by SMAC
4. Heparin binding by HAC
5. Self association by AC-SINS at pH 7.4 and pH 6.0
6. Thermostability by nanoDSF and SYPRO-DSF

Full Datasheet

Our full datasheet in Excel format (GDPa2.1.xlsx) contains the following information:

• Definitions of column headers in other datasheets
• Antibody sequences
• Assay data in "tidy data" format with one row per replicate
• Assay data summary statistics with average, standard deviation, and replicates for each assay
• Versioning

Contact

For more information on this data, see our website at https://datapoints.ginkgo.bio/, or contact us at datapoints@ginkgobioworks.com for specific questions about the data.