Masking, maintenance and mimicry: the interplay of cell-intrinsic and cell-extrinsic effects in evolutionary games
The temporal evolution of mutating pathogens in disease contexts arises from both the intrinsic properties of each subpopulation and the interactions among them, yet experimental inference often neglects the latter. Drug development studies commonly estimate selective advantages by comparing growth rates in monoculture, and in vitro monoculture dose-response curves are frequently used to justify or halt further investigation of a drug. Although many ecological models distinguish intrinsic from interaction-dependent growth rates, we show that simpler evolutionary game theory (EGT) frameworks can also be used to disentangle these contributions. We present a game-theoretic reparameterization of the replicator equation payoff matrix that separates intrinsic effects from interaction-driven contributions to frequency-dependent fitness. We also introduce an interaction-selection plot that facilitates the interpretation of the relative importance of between-population interactions compared with intrinsic evolutionary trade-offs. Using this framework, we map how interactions can mask, mirror, maintain, or mimic frequency-independent selection. We derive analytical conditions for these behaviors in both deterministic (replicator equation) and stochastic (Fokker-Planck-Kolmogorov) models, showing that simple conditions persist when mutation and noise are introduced. We validate these predictions using Wright-Fisher simulations. Applying our framework to published microbial and cancer co-culture data, we find that real systems span regimes dominated by either autonomous selection or interaction-driven effects, with interactions sometimes reversing or neutralizing frequency-independent fitness differences. Together, our results show that frequency-dependent effects can shape evolutionary dynamics in subtle and non-obvious ways, highlighting the importance of accounting for interactions when inferring fitness and predicting evolutionary outcomes.