Understanding cell state transitions and purposefully controlling them is a longstanding challenge in biology. Here, we present cell State Transition Assessment and Regulation (cSTAR), an approach to map cell states, model transitions between them, and predict targeted interventions to convert cell fate decisions. cSTAR uses omics data as input, classifies cell states, and develops a workflow that transforms the input data into mechanistic models that identify a core signaling network, which controls cell fate transitions by influencing whole-cell networks. By integrating signaling and phenotypic data, cSTAR models how cells maneuver in Waddington's landscape 1 and make decisions about which cell fate to adopt. Importantly, cSTAR devises interventions to control the movement of cells in Waddington's landscape. Testing cSTAR in a cellular model of differentiation and proliferation shows a high correlation between quantitative predictions and experimental data. Applying cSTAR to different types of perturbation and omics datasets including single cell data demonstrates its flexibility and scalability and provides new biological insights. The ability of cSTAR to identify targeted perturbations that interconvert cell fates will allow designer approaches for manipulating cellular development pathways and mechanistically underpinned therapeutic interventions.The concept of cell states is a useful lens to view and understand the organization of tissues and organisms, their development, and responses to exogenous and endogenous changes. While initially based on phenotypical descriptions, global analysis methods now can connect phenotypes with underlying molecular processes. These methods characterize cell states with fine molecular resolution and open the door to understand how cell states can evolve and transition into each other. In 1940, Waddington suggested that cells move through a landscape of mountains and valleys as rolling marbles from one (meta)stable state to another 1 . This now famous model appeals through its intuitive nature but leaves open why the marbles roll into certain valleys and whether they can revert to an initial state. Recent efforts have applied computational models to understand cell state transitions, generated by

Clear cell renal cell carcinoma (ccRCC) is the most common kidney cancer and is often caused by mutations in the oxygen-sensing machinery of kidney epithelial cells. Due to its pseudo-hypoxic state, ccRCC recruits extensive vasculature and other stromal components. Conventional cell culture methods provide poor representation of stromal cell types in primary cultures of ccRCC, and we hypothesized that mimicking the extracellular environment of the tumor would promote growth of both tumor and stromal cells. We employed proteomics to identify the components of ccRCC extracellular matrix (ECM) and found that in contrast to healthy kidney cortex, laminin, collagen IV, and entactin/nidogen are minor contributors. Instead, the ccRCC ECM is composed largely of collagen VI, fibronectin, and tenascin C. Analysis of single cell expression data indicates that cancer-associated fibroblasts are a major source of tumor ECM production. Tumor cells as well as stromal cells bind efficiently to a nine-component ECM blend characteristic of ccRCC. Primary patient-derived tumor cells bind the nine-component blend efficiently, allowing to us to establish mixed primary cultures of tumor cells and stromal cells. These miniature patient-specific replicas are conducive to microscopy and can be used to analyze interactions between cells in a model tumor microenvironment.

Ras is a key switch controlling cell behavior. In the GTP-bound form, Ras interacts with numerous effectors in a mutually exclusive manner, where individual Ras–effectors are likely part of larger cellular (sub)complexes. The molecular details of these (sub)complexes and their alteration in specific contexts are not understood. Focusing on KRAS, we performed affinity purification (AP)–mass spectrometry (MS) experiments of exogenously expressed FLAG-KRAS WT and three oncogenic mutants (“genetic contexts”) in the human Caco-2 cell line, each exposed to 11 different culture media (“culture contexts”) that mimic conditions relevant in the colon and colorectal cancer. We identified four effectors present in complex with KRAS in all genetic and growth contexts (“context-general effectors”). Seven effectors are found in KRAS complexes in only some contexts (“context-specific effectors”). Analyzing all interactors in complex with KRAS per condition, we find that the culture contexts had a larger impact on interaction rewiring than genetic contexts. We investigated how changes in the interactome impact functional outcomes and created a Shiny app for interactive visualization. We validated some of the functional differences in metabolism and proliferation. Finally, we used networks to evaluate how KRAS–effectors are involved in the modulation of functions by random walk analyses of effector-mediated (sub)complexes. Altogether, our work shows the impact of environmental contexts on network rewiring, which provides insights into tissue-specific signaling mechanisms. This may also explain why KRAS oncogenic mutants may be causing cancer only in specific tissues despite KRAS being expressed in most cells and tissues.