To expedite new medicines to patients, the biopharmaceutical industry is focusing on platform technologies and prior knowledge. The application in analytical development is not a one-size-fits-all approach, but a dynamic strategy shaped by the unique historical wealth of knowledge tied to each method at each company. Beyond this, streamlining processes plays an integral role in optimizing operational efficiency. This presentation discusses how MacroGenics strategically employs these approaches to accelerate analytical development.

The advances of new therapeutic modalities in the pharmaceutical industry drive the development of liquid chromatography (LC)-mass spectrometry (MS)-based Multi-Attribute Method (MAM). MAM has successfully demonstrated its capability in replacing some of the traditional chromatographic and electrophoretic testing methods for monitoring product quality attributes for both release and in-process testing. We list several hurdles encountered along the way of MAM adoption and discuss the new approaches to overcome them based on the latest development efforts.

Successful biologic drug characterization demands meticulous identification and control of process impurities. This presentation will explore the development and implementation of a robust platform validation strategy for process impurities. We'll discuss critical parameters, analytical techniques, and best practices for ensuring comprehensive validation. Attendees will gain insights to streamline impurity characterization, enhance product safety, and meet regulatory requirements.

Many drug product manufacturing processes require characterization of microparticulate products and contaminants. Machine learning analyses of flow imaging microscopy datasets can be used for these applications, including monitoring cell health and debris during manufacture of cell-based therapies, detection of particulate matter formed during processing of adjuvanted vaccine suspensions, and exploration of root-causes for protein aggregation. We will discuss advances in unsupervised and supervised machine learning for these analytical tasks.

The use of automation provides significant opportunities in biologics analytical development to enable increased efficiency and improved experimental design. At Merck, we have strategically positioned a group of automation experts with an explicit goal to develop user-friendly tools, templates, and designs that are shared with our scientists, including simple bench-top platforms and larger liquid handling systems. We discuss implementation of this strategy that has resulted in an automation-first mindset.

For viral vectors used in gene therapies, monitoring concentration throughout manufacture is vital for product consistency and quality. Using automation to assess AAV capsid titer enhances throughput while reducing assay hands-on time. The following presentation describes the development of an AAV capsid titer assay that combines two forms of automation: automated sample preparation and automated immunoassay and analysis. The resulting assay produces high-throughput, accurate sample results while reducing hands-on time.

Semi-automation is a proactive approach to assay development that ensures the entire process is aligned with assay requirements and is automation-compatible. Semi-automated assays are inherently designed with miniaturization and optimization considerations, ensuring seamless scalability and efficiency. This approach prevents redevelopment of assays that are incompatible with automation and greatly simplifies the process of transitioning them into fully automated systems.

Cell enumeration is vital to determine the appropriate dosage of a cell-based therapy manufacturing process. Gold standard instruments require high capital costs, maintenance, and significant sample processing for its utility. Here, I will present our recently developed portable, 3D printed, florescence microscope capable of imaging and enumerating cells in a completely automated way at the point-of-care.

At the interface between discovery and CMC, the developability team at GSK characterizes lead molecules to identify stable molecules for progression. Making use of significant investment in protein production facilities, we are now collecting biophysical data at a higher throughput and at near formulation strength. This talk will focus on how these rich data are used to improve candidate quality as well as to feed predictive models.

Protein therapeutics contain heterogenous product variants, often due to post-translational modifications (PTM). A specific PTM’s criticality depends on its potential impact to a therapeutic's efficacy and safety, which is traditionally evaluated empirically. With the burgeoning rise in computational power and biological structure elucidation, we explore the use of in silico biophysical modeling—specifically thermodynamic integration for relative binding free energies—to inform functional impacts of PTMs.

As scientists, we have expected electronic data to deliver us, only to find we are subjugated by it. The time has come to realize the unfulfilled promise. New technologies like NoSQL, ontologies, and AI/ML are rushing towards us and have outmoded our current platforms. We will explore how these technologies are changing our laboratories and increasing the value we can bring to our organizations through streamlined information delivery.

The in silico methods of identifying three-dimensional protein structure from primary sequence data is an active area of research with applications in drug discovery and development. Large-molecule proteins play crucial roles in biological processes and are key targets for drug discovery. However, predicting their complex structures accurately remains a challenging task. We discuss how ensemble modeling, which combines the predictive power of multiple models, can enhance the accuracy and reliability of protein structure predictions in the absence of experimental results.

The development of biologics generally spans years, and we build on our past experiences as we advance. As we start working on our BLA filing, we find ourselves floating in a sea of data with no clear direction, and often with databases that are only partially searchable. How do we assimilate our big data and how do we assure that future data will be better organized and more searchable?

CHO cell biopharmaceutical production faces challenges due to the demand for high-titer and complex molecules. The genome-scale metabolic model serves as a powerful tool for exploring cellular physiology and predicting cellular behaviors. By integrating omics data and advanced computational techniques, it guides metabolic engineering strategies for bioprocess optimization. Moreover, the model will guide in-process analytical testing strategies ensuring consistent product quality across all stages of production.

Analytics represent a fundamental pillar for development of biosimilars and innovative biologics. Although they are both biopharmaceuticals, the analytical strategies diverge in terms of purpose, scope, methods, and timelines. Novartis has years of experience in development of biologics and has recently transitioned into a fully innovative medicines-focused company. This presentation aims to delineate the critical distinctions between biosimilars and innovative biologics, underscoring the scientific and organizational aspects of analytical development.

Monoclonality is expected for a biologics-producing cell line. Retrospective analysis of clonality using Southern blot raised questions about clonality vs. genetic plasticity of a cell line. Long-read sequencing is an innovative assay for plasmid integration structure analysis. CRISPR/Cas9-targeted Nanopore long-read sequencing provided accurate information on the integration structure, and helped solve the clonality vs. plasticity issue. Using Southern, Sanger sequencing, and NGS as orthogonal assays confirmed the conclusions.

A 96-well plate format immuno-µPlaque assay was developed for a viral potency test to support the development of a live-attenuated quadrivalent dengue vaccine. Full automation of the assay via an integrated robotic system illustrated the potential of high-throughput cell-based analytics in the vaccine development space. A deep learning-based plaque-counting algorithm further accelerates the assay by providing analysts with precise analysis results and robust workflow.