Project
EF6-1
Heterogeneous Data Integration to Infer SARS-CoV-2 Variant-Specific Immunity for Risk Assessment and Vaccine Design
Project Heads
Max von Kleist, Claudia Schillings, René Henrion
Project Members
Carolina Barata, N. Alexia Raharinirina
Project Duration
01.01.2024 – 31.12.2025
Located at
FU Berlin
Description
SARS-CoV-2 continues to spread at alarming rates with 1-5% of the population infected at any time. Therefore, monitoring the pandemic and in particular the virus’ evolution is crucial, to identify threatening variants and to develop adapted vaccines that protect vulnerable groups or heavily-exposed individuals from emerging variants. However, despite extremely rich data sources being available for SARS-CoV-2, there has been relatively little progress in combining the available data to study mechanisms driving viral evolution, inform variant risk-assessment and vaccine design. A sensible theory is that the spread and evolution of SARS-CoV-2 is nowadays driven by the relative abundance of variant-specific susceptibiles, which is determined by infection history and the ability of infection-induced immune responses to cross-neutralize variants.
In project EF6-1, we will establish an integrative model of SARS-CoV-2 population immunity, that reconstructs infection history and incorporates immune waning and cross-neutralization between variants to estimate time-dependent variant-sspecific transmission fitness. Furthermore, we will estimate parameter and state uncertainty and estimate robust controls to suggest multi-valent vaccines that robustly protect against infection from circulating variants.
Mathematically challenges include heterogeneous data integration to construct a mechanistic and predictive virus fitness model. With regards to uncertainty estimation, we will develop sampling techniques using homotopy and preconditioning and tailor the approaches for the specific application. Robust controls (vaccine candidates) will be infered by probabilistic optimization. Here, we will focus on a infinite (e.g., indexed by continuous time as in our model) system of random inequalities. We will focus on a structural analysis of the optimization problem, which should lead us to an appropriate adaptation of efficient numerical approaches to distributions identified in the uncertainty quantification step.
Project Webpages
Selected Publications
SARS-CoV-2 Evolution on a Dynamic Immune Landscape
N. Alexia Raharinirina*, Nils Gubela*, Daniela Börnigen*, Maureen Rebecca Smith, Djin-Ye Oh, Matthias Budt, Christin Mache, Claudia Schillings, Stephan Fuchs, Ralf Dürrwald, Thorsten Wolff, Martin Hölzer, Sofia Paraskevopoulou, and Max von Kleist. (submitted)
Selected Pictures