Interactive Discussions

Engage in in-depth discussions with industry experts and your peers about the progress, trends and challenges you face in your research! Interactive discussion groups play an integral role in networking with potential collaborators, provide an opportunity to share examples from your work, and be part of a group problem-solving endeavor.

Thursday, 19 November | 15:35 – 16:15

Machine Learning for Protein Engineering Part 2

TABLE: Appropriate Use of Agentic AI in the Lab
Moderators:
M. Frank Erasmus, PhD, Head, Bioinformatics, Specifica, an IQVIA business
Monica L. Fernandez-Quintero, PhD, Associate Professor, Department of & Microbiology and Immunology, Novo Nordisk Foundation Initiative for Vaccines and Immunity (NIVI)

  • Where can agentic AI meaningfully accelerate research?
  • How should and can AI-generated outputs be verified, and what level of human intervention is required?
  • How do we protect unpublished data, ensure reproducibility, and maintain scientific standards?
  • What tools, workflows, training, and policies are needed for responsible incorporation in academic environments?
  • How does increasing autonomous AI systems change the way research is conducted, and where should we step in?
  • Who carries the liability when an agentic workflow picks a candidate that fails downstream - vendor, CRO, sponsor, or the scientist who approved it? (Most indemnification clauses today don't answer this.)
  • How do CROs and pharma prove client-data isolation when agentic tools share infrastructure, caches, and external model APIs across projects?
  • Where is the practical sandbox limit today, e.g., read-only analysis, write to ELN/LIMS, order reagents, trigger experiments, and who decides when that limit moves? In closed-loop DBTL workflows (agent designs → synthesis order → expression → assay → next round), where is the human checkpoint, and what is the maximum number of autonomous cycles before mandatory review?

TABLE: An AI-Powered Biofoundry for Protein Discovery and Engineering
Moderator: Huimin Zhao, PhD, Steven L. Miller Chair Professor, University of Illinois Urbana Champaign

  • AI-powered protein discovery and engineering: Where are current AI models most effective (e.g., structure prediction, function prediction, sequence generation), and what key limitations remain for designing proteins with novel functions?
  • Infrastructure and interoperability: What software, hardware, and data-management frameworks are required to enable seamless interaction between AI platforms, laboratory automation systems, and biofoundry operations?
  • Economic impact and industrial adoption: Which applications (therapeutics, enzymes, diagnostics, industrial biotechnology, etc.) are most likely to benefit from AI-powered biofoundries in the next 3–5 years?
  • Workforce and collaboration models: What new skills, team structures, and partnerships are needed to effectively combine AI, synthetic biology, automation engineering, and domain biology expertise?

Antibodies Against Challenging Targets

TABLE: Brain Delivery of Complex Biotherapeutics: Scientific Challenges and Future Opportunities
Moderator: Cathal Mahon, PhD, Associate Director, Protein Technologies, Denali Therapeutics Inc.

  • Overcoming biological barriers beyond BBB transport
  • Engineering biologics for optimal transport, trafficking, and cell-specific delivery
  • Translational models and biomarkers that best predict clinical success
  • Lessons learned in optimizing receptor affinity, avidity, valency, and transport efficiency
  • Future opportunities and emerging technologies for CNS drug delivery


TABLE: How is AI Facilitating the Discovery of Functional Antibodies Against Membrane Targets?
Moderator: David Felix, Team Lead, Antibody Discovery, Confo Therapeutics.

  • AI early discovery vs classical immune/synthetic repertoire screening
  • Iterative ML cycles vs wet lab selections and screening
  • PK and developability of AI-derived antibodies

Engineering the Next Generation of Bispecific Antibodies

TABLE: in vivo mRNA-Encoded T Cell Engagers
Moderator: Wei Xu, PhD, CSO, METiS TechBio

  • What are the key advantages of in vivo mRNA-encoded TCEs versus recombinant protein TCEs?
  • How might transient mRNA-driven expression alter CRS, ICANS, and on-target/off-tumor toxicity risks?
  • What are the major barriers to successful in vivo TCE development: LNP delivery, biodistribution, expression control, immunogenicity, or manufacturability?
  • Five years from now, will in vivo mRNA-encoded TCEs complement conventional bispecific antibodies—or fundamentally redefine how T-cell engagers are developed and delivered?

Protein Stability and Formation

TABLE: High-Dose Options for the Diverse Biologics Modalities and the Unknowns
Moderator: Karoline B. Bechtold-Peters, PhD, Director, Science & Technology, Drug Product Development Biologics, Novartis Pharma AG

  • Why do we need higher doses and for which (biologics and hybrid) modalities and indications?
  • What technical options are available for achieving higher doses (from formulation to devices and both)?
  • What are the pros and cons and limitations for the options? How mature are the options?
  • What are the unknowns that need to be explored? Does it make sense to teaming as industry e.g. via consortia?
  • What are regulatory trends here?

TABLE: Novel AI Architecture for Protein Stability Design
Moderator: Benjamin J. Hackel, PhD, Professor, Chemical Engineering & Materials Science, University of Minnesota
A multitude of platforms have emerged for computational design of protein developability. This roundtable will discuss factors that will drive development of improved platforms and enable optimal use of available tools.

  • Compare and contrast current platforms with rigorous benchmarking.
  • Ideate new approaches including hybrids of emerging tools, both computational and experimental.
  • Identify best practices for training and evaluation.
  • Discuss most impactful opportunities for experimental data.?

Novel Peptide Drug Discovery

TABLE: Strategies and Challenges for Intracellular Delivery of Peptides and Proteins
Moderator: Chris Alabi, PhD, Fred H. Rhodes Professor of Engineering, Cornell University

  • Delivery platform landscape; CPPs, LNPs, stapling, macrocycle, etc (strengths and weaknesses and where each is best suited)
  • Measuring and quantifying delivery (cytosolic vs endosomaly trapped, different readouts)
  • Cargo design considerations (how stability, charge, size, etc influence delivery efficiency and downstream biological activity)
  • Future directions/emerging technologies (LNPs, AI-guided delivery vehicle design, etc)