Main Conference Day 1Tuesday, 10 June 2025 - BST (British Summer Time, GMT+1)

The cytosolic Fc receptor TRIM21 uses antibodies to target proteins for degradation inside the cell. This activity provides natural protection against viral infection and has been harnessed in the targeted protein degradation technology ‘Trim-Away’. In my talk I will discuss our recent work on the molecular mechanism of cytosolic antibody-mediated degradation and how we are exploiting this mechanistic understanding to develop TRIM21-based therapeutics.

IgG4 antibodies have the unique ability to undergo Fab-arm exchange: a process that results in functional monovalent and bispecific antibodies. I will provide an overview on how Fab-arm exchange and consequent antibody valency, influences the pathogenicity of MuSK IgG4 autoantibodies. I will furthermore highlight how this insight has led to the development of a novel first-in-class treatment for neuromuscular disorders suffering from impaired neuromuscular junctions.

Radiolabeling antibodies with gamma- or positron-emitting radionuclides for imaging, or alpha- or beta-emitters for therapeutic radionuclide delivery, provides molecularly targeted approaches for diagnosis and therapy. Protein engineering has been used to optimize many characteristics of tumor-specific antibodies for immunoPET and radioimmunotherapy, particularly through PK-optimized engineered antibody fragments. Preclinical and clinical applications in radiopharmaceutical therapy of cancer and immune cell imaging demonstrate the broad potential of antibody-based theranostics.

By using combinatorial protein engineering and selection methods, proteins can be engineered and equipped with various functions. We are utilizing small, well characterized domains designed for purification, detection and therapy purposes. By tailoring the domains differently, the half-life as well as cell internalization can be manipulated to suit the intended use more efficiently. Here, the development, evaluation and use of these affinity domains in diagnostics and therapy will be discussed.

End-to-end antibody design and progression through discovery pipeline into clinically relevant context. Our extensive range of phage display libraries generate candidates with exceptional binding affinity, cross-species compatibility, and enhanced drug-like properties. Downstream evaluation of target specificity and safety assessment utilizes the Retrogenix Cell Microarray, with extensive functional biology plus PDX/CDX in vivo systems from the Charles River Tumor Model Compendium facilitating translation of preclinical candidates into clinically significant setting.

Structure-based machine learning was leveraged to design predicted antibody binders against the GPCR target C5aR1. The CDR sequences of these predicted binders were then used to construct a de-novo phage-display library. Twist’s oligo-synthesis platform enabled the fabrication of a highly diverse CDR-shuffle library with excellent variant representation and with no unwanted bias or motifs. Following panning, several high-affinity leads were identified that functionally blocked C5aR1 signalling in cellular assays.

The classic drug development funnel for mAb-like protein therapeutics starts with thousands of binders derived from a discovery engine, and each subsequent developability assay reduces the lead pipeline until only a handful winners (hopefully) are left standing. Instead, ATUM's developability engineering approach relies on utilizing information-rich multidimensional testing of a modest number of lead variants. Systematic design of the variants enable the identification and characterization of causal vs simply correlating sequence-function information. The resulting data not only dictates the 'best' solution in the searched space, but also provides boundaries for developability attributes.

We will discuss AbDiffuser, our latest equivariant and physics-informed diffusion model for the joint generation of antibody 3D structures and sequences. This will include general best practices we propose for development of such diffusion models. Laboratory experiments confirm that all 16 HER2 antibodies discovered using AbDiffuser were expressed at high levels and that 57.1% of selected designs were tight binders.

De novo design methods promise a cheaper and faster route to antibody discovery, while enabling the targeting of predetermined epitopes and the parallel screening of multiple biophysical properties. I will present recent advances to design antibodies targeting structured epitopes, to humanise nanobodies, and to simultaneously increase stability and solubility.

The emergence of ML-enabled technology platforms that aim to enhance molecule performance have the potential to revolutionize the way we approach drug discovery. However, without a purpose-built tech stack that puts data quality at the heart, many are destined to fail. This talk will focus on the deep integration of predictive assays, data generation, data capturing, and data pre-processing needed to enable iterative active learning cycles for lead optimization.

There are a plethora of novel ML protocols released, but an effort is still required to prove their benefit in therapeutic workflows. We develop and benchmark all the staple computational/ML protocols into integrated workflows, ranging from database compilation, NGS analysis, structural modeling, docking, developability, deimmunization, epitope/paratope prediction and more. We will present case studies of added benefit in bringing certain methods together as well as areas where ML underperforms.