Translational Immunoinformatics

The immune system, primarily the adaptive immunity, plays a significant role in many disease etiologies including autoimmune diseases, cancer, and infectious diseases. In our lab, we develop and apply methods from bioinformatics, machine learning, and combinatorial optimization for understanding the immune system and aiding the design of new immunotherapies. 

The “Biotherapeutics Engineering” subgroup focuses on studying the most crucial steps in the adaptive immune response, namely epitope presentation by the Major Histocompatibility Complex (MHC) and its recognition by B- and T-cell receptors. We then leverage our new insights in these areas to design novel therapeutics for cancer, infectious and auto-immune diseases.

Some notable projects include the design of chimeric antigen receptors to target self-reactive T cells, de-immunization, optimization of neoepitope-based vaccines for cancer, analyzing the T-cell repertoire from multi-omics data, and predicting epitope cleavage, MHC binding, and receptor specificity.

We do all of this and more by leveraging the latest advances in deep learning such as sequence models, multi-modal integration, multiple-instance learning and positive unlabeled learning to develop state-of-the-art computational tools to predict whether a certain peptide will trigger an immune response.

The subgroup “Biomedical Data Analysis” is focused on solving key challenges in the biomedical field through the development of state-of-the-art machine learning and biostatistics approaches. We develop Deep Learning methods for imaging and video analysis to improve health monitoring and disease diagnosis. Furthermore, we apply Machine Learning approaches to integrate molecular and clinical data to in-depth characterize disease etiologies to improve early diagnosis and provide the necessary information for effective treatments. To organize and retrieve biomedical data, we also develop smart management solutions enabling easy access and interoperability of data obtained from different studies to the benefit of the entire scientific community.

The subgroup for “Single-cell Analysis” is interested in solving major challenges in the bio-medical field by studying multimodal single-cell data integration. As the basis of most projects, we are analyse the immune response through the cells’ gene expression to understand the underlying condition. Furthermore, we complement these data with additional modalities such as chromatin accessibility, immune receptor sequences, and protein abundance information. Some of notable projects involve research on autoimmune reactions (Type 1 Diabetes), infectious diseases (SARS-CoV-2), and tumors (CAR therapies) to observe the immune response during disease progression, vaccination, or immunotherapies.

Beyond analysis, we are developing methods to integrate multimodal data efficiently, which allow us to uncover interdependencies hidden when analysing the different omics layers independently. While applying advanced statistical or machine learning techniques, we focus on reliability and reproducibility. Selected projects include learning mappings between single-cell intracellular proteome and transcriptome data from unpaired measurements, or integrating T cell receptor and gene expression data.