Products
Find the right model or tool for your drug discovery work.
AI/ML models and tools available now on TuneLab
AbLab v1, v2
Predicts antibody developability across 22 endpoints, including stability, aggregation, and solubility, from sequence alone.
Antibody-Antigen Binding Affinity (ΔΔG)
TuneLab’s Antibody–Antigen Binding Affinity model predicts how a specific mutation changes binding affinity between an antibody and its antigen.
ChemLab v1, v2 Federated
Predicts ADMET and physicochemical properties to help prioritize compounds for synthesis.
Immunogenicity Model (MHC-II Peptide Presentation)
Predicts MHC class II presentation likelihood and 9-mer binding cores for antibody peptides across 12 HLA-DRB1 alleles.
in vivo PK (developed by insitro)
Predicts preclinical in vivo PK, including clearance and Vss, across multi species, developed by insitro for Lilly.
NanoLab (VHH Developability)
Predicts five biophysical developability endpoints for VHH nanobody sequences to rank leads before wet lab testing.
Small Molecule Computational Workflows
Configurable workflows that generate, filter, score, predict properties and rank small molecules to automate drug discovery campaigns.
AbLab is Lilly TuneLab's multi-task antibody developability model. The model predicts a broad panel of biophysical developability endpoints — thermal stability, aggregation, self-interaction, hydrophobicity, non-specific binding, purity and solubility among them — directly from heavy- and light-chain sequence, with no structure required.
It is trained on Lilly's internal monoclonal antibody discovery data together with intentionally diversified antibody panels expressed and profiled at contract research organizations. It is built for antibody engineers who need to triage candidates on developability risk before committing to expression and characterization. In June, we launched AbLab v1 (version 1) with 18 endpoints, and now with this release, we have also launched AbLab v2 (version 2) with 22 endpoints. The AbLab v2 model now encodes the antibody constant region, which drives most of the performance gains on thermal stability. The sequence representation moved from a protein language model to lightweight biophysical features, which is faster and federated-training friendly. And new endpoints were added, including an expanded CRO panel and a dendritic cell internalization index.
Predictions across 22 deployed developability endpoints spanning thermal stability, aggregation, self-interaction, hydrophobicity, polyreactivity, purity, solubility and immunogenicity risk — from sequence alone, with no structure required.
A conformal prediction interval on every endpoint alongside the point estimate, which is designed to contain, on average, the true label 90% of the time Intervals are calibrated per endpoint against a held-out split.
Improved accuracy over the previous release on every deployed endpoint, with the largest gains in thermal stability. A principal driver is that it reads the five constant-domain compartments — CH1, hinge, CH2, CH3 and CL — and embeds them alongside the variable domain, which the previous version did not; thermal unfolding transitions depend heavily on constant-domain sequence.
New endpoints in this release, including an expanded set of CRO-generated tasks and a dendritic cell internalization index derived from internal data, which reports on immunogenicity risk and target-independent clearance.
A lightweight per-residue representation based on biophysical features in place of the protein language model (PLM) used previously — matching or exceeding its accuracy while running considerably faster and remaining amenable to federated training across members.
Fine tuning is available for AbLab. Fine-tuning lets a member adapt the model to their own project. It builds on the shared learnings from the federated datasets, so a member can get useful predictions from a small dataset of their own. Fine-tuned models are private to the member.
- ¹Wen Y, Cahya S, Zeng W, et al. Development of a FRET-Based Assay for Analysis of mAbs Internalization and Processing by Dendritic Cells in Preclinical Immunogenicity Risk Assessment. The AAPS Journal 22, 68 (2020). doi:10.1208/s12248-020-00444-1