Research Highlights

Research Highlights

The complexity of natural and manmade systems constantly challenges the capabilities and the robustness of computational methods, and it continues to inspire development in Computational Science. The MOSAIC group focuses on developing and applying computational methods for data analysis, simulation, and optimization of complex real-world systems. Our research builds on and extends key discoveries we have made in the past, including:


  • DC-PSE: A generalization of finite-difference and particle methods that enables fully consistent approximation of linear differential operators on (almost) arbitrary distributions of discretization points.

    B. Schrader, S. Reboux, and I. F. Sbalzarini. Discretization correction of general integral PSE operators in particle methods. J. Comput. Phys., 229:4159–4182, 2010. (PDF)

    DC-PSE enables Lagrangian simulations methods with fully runtime-adaptive resolution adaptation:

    S. Reboux, B. Schrader, and I. F. Sbalzarini. A self-organizing Lagrangian particle method for adaptive-resolution advection–diffusion simulations. J. Comput. Phys., 231:3623–3646, 2012. (PDF)

  • APR: The adaptive particle representation for optimal sampling of functions in space given an approximation error bound. It retains the good convergence properties of wavelets, but admits more efficient (linear-time) algorithms.

    B. L. Cheeseman, U. Günther, K. Gonciarz, M. Susik, and I. F. Sbalzarini. Adaptive particle representation of fluorescence microscopy images. Nat. Commun., 9:5160, 2018. (Journal Link (Open Access), PDF, Supplement PDF, Software Download, Supplementary Movies)

  • OpenFPM: A portable, scalable, and open software framework for scientific computing. It supports rapid code development and scales from single cores over multi-cores to distributed-memory clusters and GPUs.

    P. Incardona, A. Leo, Y. Zaluzhnyi, R. Ramaswamy, and I. F. Sbalzarini. OpenFPM: A scalable open framework for particle and particle-mesh codes on parallel computers. Comput. Phys. Commun., 241:155– 177, 2019. (Journal Link (Open Access), PDF, Software Download)

    OpenFPM is a modern C++ successor to the classic PPM (Parallel Particle Mesh) Library.

    I. F. Sbalzarini, J. H. Walther, M. Bergdorf, S. E. Hieber, E. M. Kotsalis, and P. Koumoutsakos. PPM – A Highly Efficient Parallel Particle-Mesh Library for the Simulation of Continuum Systems, Journal of Computational Physics 215(2):566-588, 2006. (PDF, Software Download)

    It is based on the same abstract data types and operators for distributed computing.

    I. F. Sbalzarini. Abstractions and middleware for petascale computing and beyond. Intl. J. Distr. Systems & Technol., 1(2):40–56, 2010. (PDF)

  • Partial-Propensity Methods: An efficient formulation of exact stochastic simulation algorithms for the chemical master equation, achieving linear runtime complexity for strongly coupled networks and constant complexity for weakly coupled networks of elementary reactions.

    R. Ramaswamy, N. González-Segredo, and I. F. Sbalzarini. A new class of highly efficient exact stochastic simulation algorithms for chemical reaction networks. J. Chem. Phys., 130(24):244104, 2009. (PDF, Software Download, Notes and Corrections)

    R. Ramaswamy and I. F. Sbalzarini. A partial-propensity variant of the composition-rejection stochastic simulation algorithm for chemical reaction networks. J. Chem. Phys., 132(4):044102, 2010. (PDF, Software Download)

On the basis of these key discoveries, we address applications in:


  • Active matter simulations: Numerical simulation of biological systems such as cells and tissues must account for their unique material properties. They are soft materials able to move and deform by themselves. The physics of such materials is described by the theory of active polar gels, which has been exceedingly challenging to solve numerically. We develop simulation algorithms for active matter in 2D, 3D, and on moving and deforming surfaces. This enables us to study the self-organized emergence of shape in such systems to model morphogenesis.
  • Bioimage Informatics: Many of the modern experimental assays in the life sciences deliver data in the form of digital videos or images, rather than direct quantitative measurements. Live cell imaging has become a standard method in many areas of biology. The acquired images are, however, complex and under-explored. Robust and accurate automated image processing algorithms are required for unbiased, reproducible, and quantitative analysis of the large amounts of image data acquired. In addition, it is imperative for subsequent modeling that the confidence and reliability of the image processing results are known.
  • Data-driven modeling: Complementing modeling based on physical principles, we use experimental data to infer and optimize predictive spatiotemporal models of biological systems. We develop and apply methods to learn partial differential equation (PDE) models from microscopy images and videos, to robustly estimate unknown model parameters from data, and to complement bottom-up numerical simulations with data-driven surrogate models. This is particularly important in developmental biology, where consensus physics models are often not available, or the underlying physics may not be entirely known.

More Previous Contributions

Some additional previous contributions of ours include:


  • an efficient algorithm for single-particle tracking without motion priors [1] and its application to virus entry [2,3].
  • a trainable algorithm for detecting and extracting motion patterns from trajectories of moving objects and its application to virus entry [4].
  • an image segmentation framework that accounts for and corrects the optical blur introduced by the microscope optics, allowing nanometer-precise reconstruction of the outlines of small intracellular objects [5]. It has been applied to characterize for the first time the morphodynamics of endosomes in live cells [6].
  • a new particle-based image segmentation framework that is particularly well suited for 2D and 3D fluorescence microscopy [7].
  • a statistical framework for inferring interactions between objects in images [8]. This extends classical co-localization analysis to interaction analysis and allows correcting the systematic errors inherent to co-localization analysis.
  • a new class of exact simulation algorithms for biochemical networks in space [9] and time [10-12]. The algorithms are orders of magnitude faster than previous ones at the same accuracy. This has enabled the discovery of fundamental effects in chemical kinetics [13-15].
  • a self-organizing adaptive particle method for simulating continuum models in complex and multi-scale geometries [16,17]. The number and placement of particles is automatically determined by the method, making it the most user-friendly numerical simulation scheme available to date. the first randomized optimization heuristic ever reported to robustly solve a multi-funnel problem [18] and its application to protein structures [19].
  • the PPM Library, a parallel computing middleware for hybrid particle-mesh methods [20,21]. PPM-based simulations can be implemented in a fraction of the traditional software development time (days instead of years) and often outperform hand-written simulation programs [22].
  1. I. F. Sbalzarini and P. Koumoutsakos. Feature point tracking and trajectory analysis for video imaging in cell biology. J. Struct. Biol., 151(2):182–195, 2005.

  2. H. Ewers, A. E. Smith, I. F. Sbalzarini, H. Lilie, P. Koumoutsakos, and A. Helenius. Single-particle tracking of murine polyoma virus-like particles on live cells and artificial membranes. Proc. Natl. Acad. Sci. USA, 102(42):15110–15115, 2005.

  3. Y. Yamauchi, H. Boukari, I. Banerjee, I. F. Sbalzarini, P. Horvath, and A. Helenius. Histone deacetylase 8 is required for centrosome cohesion and influenza A virus entry. PLoS Pathog., 7(10):e1002316, 2011.

  4. J. A. Helmuth, C. J. Burckhardt, P. Koumoutsakos, U. F. Greber, and I. F. Sbalzarini. A novel supervised trajectory segmentation algorithm identifies distinct types of human adenovirus motion in host cells. J. Struct. Biol., 159(3):347–358, 2007.

  5. J. A. Helmuth and I. F. Sbalzarini. Deconvolving active contours for fluorescence microscopy images. In Proc. Intl. Symp. Visual Computing (ISVC), volume 5875 of Lecture Notes in Computer Science, pages 544–553, Las Vegas, USA, November 2009. Springer.

  6. J. A. Helmuth, C. J. Burckhardt, U. F. Greber, and I. F. Sbalzarini. Shape reconstruction of subcellular structures from live cell fluorescence microscopy images. J. Struct. Biol., 167:1–10, 2009.

  7. J. Cardinale, G. Paul, and I. F. Sbalzarini. Discrete region competition for unknown numbers of connected regions. IEEE Trans. Image Process., 2012.

  8. J. A. Helmuth, G. Paul, and I. F. Sbalzarini. Beyond co-localization: inferring spatial interactions between sub-cellular structures from microscopy images. BMC Bioinformatics, 11:372, 2010.

  9. R. Ramaswamy and I. F. Sbalzarini. Exact on-lattice stochastic reaction-diffusion simulations using partial-propensity methods. J. Chem. Phys., 135:244103, 2011.

  10. R. Ramaswamy, N. González-Segredo, and I. F. Sbalzarini. A new class of highly efficient exact stochastic simulation algorithms for chemical reaction networks. J. Chem. Phys., 130(24):244104, 2009.

  11. R. Ramaswamy and I. F. Sbalzarini. A partial-propensity variant of the composition-rejection stochastic simulation algorithm for chemical reaction networks. J. Chem. Phys., 132(4):044102, 2010.

  12. R. Ramaswamy and I. F. Sbalzarini. A partial-propensity formulation of the stochastic simulation algorithm for chemical reaction networks with delays. J. Chem. Phys., 134:014106, 2011.

  13. R. Ramaswamy, N. González-Segredo, I. F. Sbalzarini, and R. Grima. Discreteness-induced concentraiton inversion in mesoscopic chemical systems. Nat. Commun., 3:779, 2012.

  14. R. Ramaswamy and I. F. Sbalzarini. Intrinsic noise alters the frequency spectrum of mesoscopic oscillatory chemical reaction systems. Sci. Rep., 1:154, 2011.

  15. R. Ramaswamy, I. F. Sbalzarini, and N. González-Segredo. Noise-induced modulation of the relaxation kinetics around a non-equilibrium steady state of non-linear chemical reaction networks. PLoS ONE, 6(1):e16045, 2011.

  16. B. Schrader, S. Reboux, and I. F. Sbalzarini. Discretization correction of general integral PSE operators in particle methods. J. Comput. Phys., 229:4159–4182, 2010.

  17. S. Reboux, B. Schrader, and I. F. Sbalzarini. A self-organizing Lagrangian particle method for adaptive-resolution advection–diffusion simulations. J. Comput. Phys., 231:3623–3646, 2012.

  18. C. L. Müller, B. Baumgartner, and I. F. Sbalzarini. Particle swarm CMA evolution strategy for the optimization of multi-funnel landscapes. In Proc. IEEE Congress on Evolutionary Computation (CEC), pages 2685–2692, Trondheim, Norway, May 2009. IEEE.

  19. C. L. Müller, I. F. Sbalzarini, W. F. van Gunsteren, B. Žagrović, and P. H. Hünenberger. In the eye of the beholder: Inhomogeneous distribution of high-resolution shapes within the random-walk ensemble. J. Chem. Phys., 130(21):214904, 2009.

  20. I. F. Sbalzarini, J. H. Walther, M. Bergdorf, S. E. Hieber, E. M. Kotsalis, and P. Koumoutsakos. PPM – a highly efficient parallel particle-mesh library for the simulation of continuum systems. J. Comput. Phys., 215(2):566–588, 2006.

  21. O. Awile, O. Demirel, and I. F. Sbalzarini. Toward an object-oriented core of the PPM library. In Proc. ICNAAM, Numerical Analysis and Applied Mathematics, International Conference, pages 1313–1316. AIP, 2010.

  22. I. F. Sbalzarini. Abstractions and middleware for petascale computing and beyond. Intl. J. Distr. Systems & Technol., 1(2):40–56, 2010.