Abstract
AbstractDecisions in biology happen fast and are driven by evolution to optimize survival chances. In platelets, this is achieved by organizing signaling cascades into rapid decision-funnels with modulatory crosstalk. We show that network decision processes underlying cellular decisions are tough to solve (equivalent to classical satisfiability problems, SAT). Hence, heuristics, modular decision-making, and decision funnels are required for efficient decisions.We establish this using a seven-layer platelet decision network that agrees well with all available genetic and functional experimental data. Platelet decision cascades are robust to perturbations: For example, receptors such as TRPM7 modulate platelet activity. However, knockouts of the receptors still leave platelets reactive overall. Dynamic control resolves relaying functions from kinases to cytoskeleton alterations. This allows fast execution of platelet shape change or aggregation. Stress conditions can shift platelet decision funnels towards constant activation of aggregation or immune signaling, causing thrombosis or thrombo-inflammation. Based on the network dynamics, we conclude that platelets pragmatically resolve the complex (non-polynomial (NP)) cellular decision problems by using a similar relaxation to those proposed in mathematics – many different configurations end up in similar states. Metamathematical considerations (no mathematical proof) suggest that NP problems are more complex then P problems.One sentence abstractWe show that cellular decision problems like the platelet signaling cascade may need unexpectedly long to solve but in general, they are efficiently solved using heuristics (“decision funnels”), implying fast decisions but the risk of chronic stress and inflammation.
Publisher
Cold Spring Harbor Laboratory