team
CoSysBio
Team manager: Niarakis Anna
Computational models and Digital Twins for complex human pathologies: Digital twins, virtual replicas representing physical objects, can potentially revolutionize the field of Biomedicine. They are an emergent concept, especially in personalized medicine (Laubenbacher, Niarakis et al., 2022; Niarakis et al., 2024). Recent advances in AI technology and the advent of omics make medical Digital Twins more promising than ever. Robust and reliable computational models spanning all biological layers, such as gene expression, signalling and metabolism could transform the way we treat Big Data for the benefit of precision medicine and improved medical care – tailored to the needs of each patient.
Computational models have a crucial role in biology, helping to test hypotheses and predict the system of interest. Data-driven immune digital twins can be constructed using machine learning or artificial intelligence algorithms to integrate several types of biological data and predict changes regarding the patient’s well-being, disease risk or even response to therapy. Patient-specific data used to construct a digital twin can vary from multi-omic data, whole genome or exome sequence or clinical data, and information on diet or lifestyle habits. However, data-driven twins obtained with machine learning or AI do not provide insights regarding the biological mechanism responsible for the changes. Mechanistic twins can model specific mechanisms within the body or the cell type, tissue, or organ of interest and can be calibrated using patient phenotype characteristics and real-time feeds. Hybrid, causal AI twins combine the best of both worlds, providing predictions and causal, mechanistic hypotheses.
Projet 1
In Diogenes, we aim to create a medical digital twin for Rheumatoid Arthritis (RA-DT), composed of a basic computational model with several cell types present in the joint, such as chondrocytes, synoviocytes and fibroblasts. The characteristic tissues present in the joint, such as cartilage, bone and synovial tissue, as well as other factors, such as oxidative stress and immune response, should be considered to build a composable, multicellular and multiscale model of the disease. The goal is to predict optimal conditions that minimize bone erosion, cartilage destruction and inflammation. The RA-DT will be developed as a multiscale model and will be exclusively confronted with published human-derived data from small-scale experiments and patient omics datasets. Then, the calibrated model will be used to predict the behavior of the system regarding the effect of targeting various factors. Our goal is to demonstrate that RA-DT can achieve similar results, or even surpass in vivo mouse models.
Projet 2
Atopic dermatitis (AD), often called eczema, is characterized by chronic skin inflammation involving close communication between mast cells (MCs) and nociceptors. The Digi-DermA project aims to (a) build a digital twin to study these interactions, more precisely the activation of MCs by various signaling cascades (canonical and non-canonical) and the respective ligands (e.g. neuropeptide). To build the model, we will use large- and small-scale data, from both mouse and human models, and we will also integrate empirical knowledge provided by experts in the field. To this end, we will (b) develop appropriate formal methods, modeling tools and simulation infrastructure to help infer and understand causality. Indeed, the modeling design will include large-scale Boolean models to capture broad and intertwined signaling cascades and smaller kinetic or agent-based models, which would allow capturing temporal aspects of disease manifestation (acute or late inflammatory responses, affinities of different neuropeptides). This modeling will help to (c) identify knowledge gaps and missing information, thus guiding the design of experimental designs to generate new discoveries and consequently improve the reliability of computational models. The predictions of the digital twin will be validated using skin samples from AD patients and data from in vivo animal models. This will lead to the (d) contextualization of the nociceptor-mast cell model to simulate local and systemic manifestations in various organs and at the body scale. Partners: Dr. Nicola Gaudenzio, INFINITY, Inserm, Toulouse; Dr. Fabien Crauste, CNRS, Paris Cité, Paris; Dr. Sylvain Soliman, CR, Inria Saclay, Saclay.
Projet 3
Computational systems biology approach to unveil molecular interactions in Sjogren’s disease (SjD) pathogenesis. SjD presents an unmet medical challenge as there is currently no cure. Despite advances in understanding the immunopathogenesis of SjD, there is still a pressing need to identify novel biomarkers and therapeutic targets, for better patient stratification and personalized treatment. We aim to create a) a fully detailed molecular interaction map (MIM) including all the signalling and molecular pathways implicated in SjD pathogenesis; b) a large-scale mechanistic model to enable in silico simulations of perturbations including drug interventions, and the generation of hypothesis-driven predictions, and c) to characterize the mechanisms of activation of lymphocytes B in Sjogren patients that develop lymphoma. Project in collaboration with the Rheumatology Service of Kremlin Bicetre Hospital, partners: Dr Gaetane Nocturne, PU-PH, Dr Xavier Mariette, PU-PH, and the NECESSITY consortium.
Projet 2
Multiscale Modeling of Inflammation. Inflammation is crucial for the body’s defense against harmful stimuli, such as infections, tissue damage, and various stressors, and is required to initiate subsequent repair and healing processes. However, dysregulated inflammation can also play a detrimental pathophysiological role in various diseases. In acute situations, excessive inflammation can lead to unwanted collateral tissue damage, leading to widespread excessive inflammation, organ dysfunction/injury, and death. Alternatively, chronic or persistent inflammation drives the development and progression of various chronic diseases, including autoimmune diseases, cardiovascular diseases, neurodegenerative diseases, and some cancers. Given its pervasive nature, understanding how to control inflammation effectively will have a profound impact on many diseases.
Consequently, focusing on specific pathological conditions associated with inflammation—where dysregulation has distinct local and systemic manifestations, cellular interactions involving immune cells are more measurable and quantifiable, and multi-omics data are available — appears a valid starting point for initial “proof of concept” Immune Digital Twins implementations. Furthermore, because inflammation is pervasive in many disease contexts, it could serve as a foundational immune response model, adaptable to a wide range of cases. Project in collaboration with the RDA Europe Working Group” Building Immune Digital Twins”.
– Zerrouk, N., Alcraft, R., Hall, B.A., Auge, F, Niarakis, A, Large-scale computational modelling of the M1 and M2 synovial macrophages in rheumatoid arthritis. npj Syst Biol Appl 10, 10 (2024). https://doi.org/10.1038/s41540-024-00337-5
– A Niarakis, M Ostaszewski, A Mazein, I Kuperstein, M Kutmon, et al., Drug-Target identification in COVID-19 disease mechanisms using computational systems biology approaches, Front. Immunol. Sec. Systems Immunology; Volume 14 – 2023 | doi: 10.3389/fimmu.2023.1282859
– Vidisha Singh, Aurelien Naldi, Sylvain Soliman, Anna Niarakis, A large-scale Boolean model of the Rheumatoid Arthritis Fibroblast-Like Synoviocytes predicts drug synergies in the arthritic joint (2022), npj Syst Biol Appl 9, 33 (2023). https://doi.org/10.1038/s41540-023-00294-5
– Sahar Aghakhani, Sacha E Silva-Saffar, Sylvain Soliman, Anna Niarakis, Hybrid computational modeling highlights reverse Warburg effect in breast cancer-associated fibroblasts, Computational and Structural Biotechnology Journal, DOI: https://doi.org/10.1016/j.csbj.2023.08.015
– Sahar Aghakhani, Sylvain Soliman, Anna Niarakis, Metabolic Reprogramming in Rheumatoid Arthritis Synovial Fibroblasts: a Hybrid Modeling Approach, PLOS Computational Biology 18 (12), e1010408 (2022) https://doi.org/10.1371/journal.pcbi.1010408
– N Zerrouk, S Aghakhani, V Singh, F Augé, A Niarakis, A Mechanistic Cellular Atlas of the Rheumatic Joint, Frontiers in Systems Biology, 2022; 2 DOI=10.3389/fsysb.2022.92579
– Anna Niarakis, Dagmar Waltemath, James Glazier, Falk Schreiber, Sarah M Keating, David Nickerson, Claudine Chaouiya, Anne Siegel, Vincent Noël, Henning Hermjakob, Tomáš Helikar, Sylvain Soliman, Laurence Calzone, Addressing barriers in comprehensiveness, accessibility, reusability, interoperability and reproducibility of computational models in systems biology, Briefings in Bioinformatics, 2022; bbac212, https://doi.org/10.1093/bib/bbac212
– R. Laubenbacher, A. Niarakis, T. Helikar, G. An, B. Shapiro, R. S. Malik-Sheriff , T.J. Sego, A. Knapp, P. Macklin, J. A. Glazier Building Digital Twins of the Human Immune System: Toward A Roadmap, npj Digit. Med. 5, 64 (2022). https://doi.org/10.1038/s41746-022-00610-z
Affiliation
