PEGS Europe Breakout Discussions

Wednesday 16 November 14:45 – 15:25

Machine Learning Approaches for Protein Engineering

Table 1: Best Practices for Using Machine Learning in NGS-Guided Antibody Discovery
Moderator: M. Frank Erasmus, PhD, Head, Bioinformatics, Specifica, Inc.

  • What questions do you aim to address within a given NGS-guided discovery campaign?
  • How does unsupervised or supervised machine learning aid in this NGS-guided discovery effort?
  • Is deep learning required for your particular application or do shallow learning approaches/simple heuristics suffice?
  • How do you collect/prepare your data to established an accurate ground truth?
  • How do you validate your ML model?
  • What data encoding/reduction methods do you employ (e.g. one-hot encoding, physicochemical tokenization) to represent your sequence data?
  • Does 3D coordinate information enhance your sequence-based dataset? Which tools do you use when structure is unavailable?


Table 2: Technologies to Discover Functional Antibodies Targeting GPCRs
Peter McNamara, PhD, Senior Vice President, Head of Research, Tectonic Therapeutics

  • Strategies employed for different GPCR subclasses
  • Antigen design and purification strategies
  • In vitro display approaches vs in vivo immunization strategies
  • Library designs for in vitro display
  • Single chain vs Fab discovery strategies
  • Discovery of biased agonists/antagonists
  • Measuring Ab kinetics on GPCRs


Table 3: Latest in the Development of Bispecific Antibodies and What’s Next on the Horizon
Moderator: Christian Klein, PhD, Cancer Immunotherapy Discovery, Roche Innovation Center Zurich, Roche Pharma Research & Early Development, pRED

Immunotherapy Safety & Efficacy

Table 4: Improving Immunotherapy Safety and Managing Toxicity
Moderator: Sara M. Mangsbo, PhD, Senior Lecturer, Pharmacy, Uppsala University

  • Impact of tumour microenvironment on safety and efficacy
  • Engineering immunotherapeutics to improve safety and reduce toxicity
  • Emerging trends and technologies

Protein Stability & Aggregation

Table 5: Assessment of Self or Non-Specific Interactions of Therapeutic Proteins and ADCs
Moderator: Bernhard Valldorf, PhD, Principal Scientist & Lab Head, Formulation Development, EMD Serono

  • In silico molecular descriptors – how well can we predict aggregation behavior of mAbs or fusion proteins? Where are the limitations?
  • What is your favorite method/tool to characterize self-association and why?
  • How does conjugation site and method influence self or non-specific interactions? How can we assess developability of ADCs?
  • What kind of short-term dataset is required to predict long-term stability?
  • High concentrated formulations – how can we prevent aggregation with new excipients or technical solutions?

Table 6: Polysorbate Challenges in Biotherapeutic Formulations
Moderator: Camille Dagallier, PhD, Senior Formulation Scientist, Biologics Drug Product Development & Manufacturing, Sanofi R&D

  • Root causes for polysorbate oxidation
  • Mitigation strategies against polysorbate oxidation
  • Alternative approaches and surfactants

Protein Purification Technologies

Table 7: Transient Protein Production Challenges
Moderator: Richard Altman, MS, Field Application Scientist, Protein Expression, Biosciences Division, Life Sciences Solutions Group, Thermo Fisher Scientific
Transient protein production in mammalian cells continues its evolution as an integral part of the biotherapeutic drug discovery process as well as an important tool to generate recombinant proteins for a variety of other applications. We discuss challenges facing researchers as they try to meet an expanding demand for transiently produced recombinant protein.

  • What are the current challenges to transient protein production?
  • How do we optimize the whole protein expression process?
  • What parameters can impact the quality or physical attributes of transiently produced proteins?

Table 8: High-Throughput (HTP) Protein Production
Moderator: Nicola Burgess-Brown, PhD, Director of Enzymology and Protein Engineering, Exact Sciences Innovation, Oxford, UK

  • Benefits of testing multiple constructs in parallel.
  • HTP cloning or gene synthesis?
  • HTP expression screening in multiple hosts: What scale, conditions, equipment, readout? How to screen intracellular, secreted and membrane proteins in HTP?
  • Choice of purification tags for HTP screening.
  • Challenges of working in HTP: What conditions to test first to increase success?