AMR Accelerator at the AMR Conference 2026: Showcasing Data, Models, and Translational Science

Advancing Data-Driven Discovery: Open Call for Permeability Data

A key highlight is the work of the Scientific Interest Group in Machine Learning of the AMR Accelerator, which is developing predictive models of compound permeability in Gram-negative bacteria. As membrane permeability remains a major barrier to antibiotic efficacy, this effort aims to enable the in silico prioritisation of compounds with favourable uptake profiles.

To support this initiative, the group has launched an open call for experimental datasets, including compound uptake, accumulation, and relevant proxy readouts such as MIC data linked to chemical structures. This collaborative effort seeks to accelerate model development while improving mechanistic understanding of compound entry into Gram-negative pathogens.

So keep an eye out for the poster titled “Tackling the Gram-Negative Permeability Barrier Through Collaborative Data-Driven Prediction” by Leonie von Berlin, Gesa Witt, Yojana Gadiya, Josepine Fernow, and Frederike Deroose. Or check out the call for data directly, follwing this link: Open Call for Data.

Strengthening TB Drug Development with HFS-TB Models

A contribution from the ERA4TB project focuses on the application of the hollow-fiber system for tuberculosis (HFS-TB), a preclinical in vitro PK/PD platform endorsed by the European Medicines Agency. This system enables simulation of human-like pharmacokinetics and supports the design of Phase II/III clinical trials.

The presented work evaluates the implementation of the next-generation diarylquinoline TBAJ-587, a promising alternative to bedaquiline with improved safety and efficacy profiles. The study highlights important experimental considerations, such as non-specific drug binding within the system, and proposes methodological adaptations to ensure accurate exposure simulation.

By integrating physiologically-based pharmacokinetic (PBPK) modelling with HFS-TB experiments, the researchers demonstrate how lung drug concentrations can be translated from preclinical models to human scenarios. The results provide valuable insights into the bactericidal activity of TBAJ-587 across replicating and non-replicating Mycobacterium tuberculosis strains, while also underlining the importance of system-specific factors in interpreting outcomes.

If you want to learn more or connect with the researchers, look for their poster titled “Assessment of feasibility of the hollow-fiber system for PKPD studies of diarylquinolines” at the AMR Conference. You find the poster abstract on the AMR Conference website.

Improving Translational Relevance of Pneumonia Models

In addition, AMR Accelerator partners will present ongoing work to enhance the predictive value of pneumonia infection models. These efforts focus on improving reproducibility and translational alignment with clinical outcomes, ultimately supporting more reliable evaluation of antibacterial therapies in preclinical settings. Learn more about the COMBINE Pneumonia Model in this recent news on our website. Get to meet the researchers behind the model at the conference by looking for their poster “The COMBINE Pneumonia model; Multicenter Confirmation of Virulence”.

Collaboration and Engagement

The AMR Conference provides an important platform for exchange across academia, industry, and public stakeholders. The AMR Accelerator welcomes discussions with researchers, data contributors, and potential collaborators interested in advancing antibiotic development through integrated, multidisciplinary approaches.

We look forward to engaging with the AMR community in Basel and sharing progress from across the programme.