Simon Syga

01 · Tumor evolution

Why tumor edges and cores evolve differently

QuestionCancer cells can switch between migrating and dividing in response to their surroundings. How does evolution tune this switch, and what does the resulting heterogeneity mean for therapy?

What we foundIn a spatial model that tracks the genotype and phenotype of every cell, cells at the invasive edge evolve to favor migration and cells in the crowded core evolve to favor proliferation. Different genetic configurations can produce the same pattern. After simulated treatment, tumors with more phenotypic heterogeneity recurred sooner, whereas more genetic heterogeneity was associated with later recurrence.

NowA new cellular automaton framework makes such genotype–phenotype models easier to build and analyze. Current collaborations apply these ideas to treatment resistance in lung cancer and to heterogeneity in glioblastoma.

PapersPLOS Comput Biol 2024 · Eur Phys J Spec Top 2026 · PhD thesis, TU Dresden 2024

02 · Cancer invasion

How tissue confinement shapes cancer invasion

QuestionBreast cancer cells invade surrounding tissue either as cohesive groups or as single cells. What decides between the two?

What we foundWith Peter Friedl’s group, we combined organotypic culture, intravital microscopy and cellular automaton simulations. Three-dimensional tissue boundaries regulate cell density and keep cells moving collectively even when cell–cell junctions are weakened; single cells escape only where free space opens up. In a follow-up study, the model predicted that cells can invade along tissue clefts without degrading the surrounding matrix, which experiments confirmed.

NowI lead the modeling in the ERC Advanced Grant subLETHAL, which extends this work to how invading cells respond to therapy.

PapersNature Cell Biology 2020 · J Cell Sci 2023

03 · Collective behavior & methods

From local rules to collective behavior

QuestionCells and animal groups coordinate their movement without a leader. Which local interaction rules produce collective behavior, and how can we build discrete models that we can analyze and trust?

What we foundBIO-LGCA, a lattice-gas cellular automaton framework I helped develop, derives interaction rules from biophysics or data and predicts collective phases such as aggregation and jamming by mean-field analysis. We showed how such lattice models can produce artificial checkerboard patterns and how to prevent them. In a recent preprint, a rule in which an agent occasionally follows a strongly deviating neighbor triggers cascades of reorientation that make swarms more responsive while keeping them cohesive.

PapersNew J Phys 2020 · PLOS Comput Biol 2021 · BIOMATH 2023 · arXiv 2026 · biolgca software

04 · Epidemics

What contact restrictions changed in COVID-19 transmission

QuestionDuring the first COVID-19 wave in Germany, non-pharmaceutical interventions changed how people met. Which features of the contact network did they change?

What we foundCombining an agent-based model on a small-world network with Bayesian inference, we found that the interventions mainly removed random, long-range contacts and increased clustering. This switched the epidemic from exponential growth to a regime of constant new cases, in which the disease spreads like a wave at a speed that mean-field theory predicts.

For a general audience, I co-authored the book Mathematik der Pandemie (Springer, 2022).

PapersScientific Reports 2021 · Mathematik der Pandemie, 2022