Ginkgo Datapoints

Ginkgo Datapoints is a Bio × AI data provider, based in Boston and part of Ginkgo Bioworks. We generate biological data at scale using validated methods, run by robots rather than by hand. Our data generation services cover three platforms:

• Antibody Developability (AbDev): screens of 2,400+ antibody designs in parallel across 12+ biophysical assays.
• Functional Biology (FunBio): industrialized DRUG-seq and Cell Painting across 80+ cell types, including primary cells, plus custom high-throughput screens.
• Small Molecule ADME: Tier 1 ADME at lead or library scale, including end-to-end service with ADME-One™.

Every data return is AI/ML-ready. Learn more at datapoints.ginkgo.bio.

🧬 Virtual Cell Pharmacology Initiative (VCPI)

VCPI is our flagship open-science project: we're building a public gene–drug response atlas of 100,000 compounds, each profiled using DRUG-seq in a 6-point dose response. The data is free to access, and open submissions are live, so you can have your own compounds profiled for free.

On Hugging Face:

• vcpi-phenotype-ranking-v1: a benchmark that asks a model to identify a compound from its DRUG-seq response in THP-1 cells. It was built on VCPI data by Mengbo Wang (Purdue University).

📊 Open datasets

Antibody developability

Transcriptomics and imaging

The Datapoints website hosts more data, including GDPsm1 (Tier 1 ADME for 320 LOPAC compounds).

🏆 AbDev Competition

The 2025 Antibody Developability Competition was the first open benchmark of antibody developability prediction, built on GDPa1 with GDPa3 as the held-out test set. The outcomes are published in mAbs.

📄 Publications

• GDPa1: PROPHET-Ab: A high-throughput platform for biophysical antibody developability assessment to enable AI/ML model training
• GDPx2: Mapping the transcriptional landscape of drug responses in primary human cells using high-throughput DRUG-seq
• GDPa3: 2025 Ginkgo Datapoints Antibody Developability Competition outcomes (mAbs, 2026)
• GDPa4: Decoding bispecific antibody developability (bioRxiv, 2026)
• GDPa5: Predicting VHH-Fc developability from large-scale IgG data (bioRxiv, 2026)

See also our blog post with Hugging Face about publishing our first datasets here.