We are establishing an integrated control strategy for undesired mis-paired variants in bispecific antibodies with variable modality complexity. Analytical approaches are selected by considering separation mode, throughput, process requirements, and pairing efficiency. The quality impact of the variants is also evaluated to define appropriate control levels. This presentation illustrates these concepts with practical examples from bispecific antibody development.

Suspension formulations comprising protein powders dispersed in non-aqueous vehicles represent a promising approach to enable high-dose subcutaneous administration of therapeutic proteins. However, robust analytical characterisation of these systems remains challenging. In this work, we evaluated multiple strategies to analyse protein quality and formulation composition of non-aqueous suspensions. Our results highlight that optimised sample preparation was critical to enable drug product characterisation with biochemical analytical techniques.

Recombinantly produced proteins carrying bacterial O-antigen glycosylation represent promising glycoconjugate vaccine candidates for the prevention of bacterial infections. Liquid chromatography–mass spectrometry (LC–MS) has become an essential tool for the comprehensive structural characterisation of their N-glycosylation profile. Here, we characterise glycoconjugates at two structural levels—glycopeptides and intact bioconjugate glycoforms—and analyse related O-polysaccharides derived from lipid-linked oligosaccharides by LC–MS.

This work presents an automated approach for analysing DSF traces that incorporates prior knowledge of the molecule under study. It automates the analysis of DSF thermograms, enabling scalable, consistent interpretation of large datasets without manual intervention. By improving the accuracy of extracted thermodynamic parameters, it allows DSF to approach the reliability of higher-resolution techniques, expanding its utility in antibody characterisation and developability assessment.

To identify effective immune engagers, including T cell therapies or T cell-dependent bispecific antibodies, robust screening of T cell-mediated tumor killing co-culture is crucial. Traditional imaging relies on fluorescently labeled cells, risking artifacts and phototoxicity. We introduce an AI/ML-based method utilizing only brightfield images to identify phenotypic changes, eliminating fluorescent markers. This innovative approach applied to T cell-killing assays maintains consistency, enhances efficiency, and allows analysis of diverse tumor cells without complex segmentation.

Antibody discovery is entering a new paradigm driven by the convergence of biology and machine intelligence. We present the MAb-Intelligence Triad, integrating natural antibody evolution from clonally expanded, somatically hypermutated B cell repertoires, human expertise in library design and directed evolution (Kling-EVOLVE), and AI models trained on large-scale paired VH–VL datasets. Using SARS-CoV-2 as a benchmark, we demonstrate prediction of improved affinity and cross-reactivity, with emerging applications in oncology and autoimmunity.

Potency assays are indispensable for biologics development and for understanding how drug quality attributes affect bioactivity. However, establishing potency assays that are robust enough for quality control often requires significant effort. In this presentation, we discuss phase-appropriate strategies—supported by examples—for selecting and developing mechanism-of-action-relevant potency assays for both standard and complex biologics.

Biophysics screens are crucial to accelerate discovery and lead optimisation as they predict developability risks. We have established an end-to-end high-throughput automation platform for miniaturised biophysical assays. Through a user-friendly interface, with a single click we can pull sample metadata, start robotic workflows, and register samples, plates, and data in internal databases and ELNs, enabling data readiness for compound screening, lead selection, formulation development, and for ML models.

The development of effective anti-fibrotic biotherapeutics is limited by a lack of human-relevant analytical frameworks linking immune modulation to functional tissue outcomes. Here, we present an immune-competent, human iPSC-based analytical platform that quantitatively resolves macrophage–fibroblast crosstalk in cardiac fibrosis. By integrating high-resolution spectral flow cytometry with functional co-culture and conditioned media assays, we define macrophage phenotypic states and directly map these states to cardiac fibroblast activation and extracellular matrix responses. This platform enables mechanism-linked potency and differentiation assessment of macrophage-targeting antibodies, supporting target validation, candidate ranking, and potency assessment for next-generation anti-fibrotic biologics.