December Q4 2025 - Product Update Webinar

We recently hosted the Q4 2025 Eurasia Product Update webinar highlighting major advancements in CDD Vault’s registration and AI modules. These enhancements were designed based on direct feedback from scientists working with complex biologics, antibody drug conjugates (ADCs), and structure-based discovery workflows. Below is a recap of the major updates and demonstrations

Enhanced Registration for Antibodies and ADCs

Large molecule programs are growing in complexity, and CDD Vault’s registration capabilities have expanded to meet this demand. The new features focus on capturing rich structural data and metadata for biologics workflows.

Key capabilities introduced:

• Support for chemically aware registration of antibodies with up to four chains, including bispecific, trispecific, and tetraspecific formats.
• Explicit definition of disulfide bonds and conjugation sites for precise structural representation.
• Ability to register antibody drug conjugates (ADCs) with up to four payload attachments, with support for site-specific or stochastic linkage.
• Integration of these large molecule entities into existing vault workflows and macrocycle libraries, improving consistency and traceability.

During the demo, our product team showed how registration operates in a three-layer paradigm:

• Layer 1 incorporates foundational monomer chemistry (including custom monomer libraries for non-standard residues).
• Layer 2 assembles component-based entities combining chains and conjugates.
• Layer 3 enriches structures with metadata (domains, bonds, hinge regions) for advanced visualization.

In practice, users were shown how antibodies and ADCs appear in search reports with associated metadata, assay links, and inventory data, treated as first-class entities alongside small molecules and peptides.

Boltz 2 AI Integration for Structure Prediction and Docking

On the AI side, the webinar introduced an integrated Boltz 2 docking workflow within CDD Vault, aimed at making structure prediction and affinity estimation part of everyday research workflows.

Highlights of Boltz 2 integration:

• Run Boltz 2 docking directly on protein-ligand pairs registered in Vault to receive predicted structures, binding affinity estimates, and binder probabilities.
• Trees and outputs from docking jobs are stored within Vault tables, allowing review and annotation alongside other project data.
• Jobs can be executed individually or at scale using import spreadsheets, enabling high-throughput prediction pipelines.

The demonstration covered:

• How to create a docking protocol with trigger and modeling readouts.
• Upload of protein targets via sequence text or SIF format (a modern format for crystallographic data).
• Submission of docking jobs from the run data tab, with interactive review of model outputs in the Mole Viewer.

Users also saw how Boltz 2 outputs include predicted pIC50 affinity values, a binary binder probability score, and confidence metrics, with visual tools for examining interactions and surfaces.