BackgroundPersonalized, precision, P4, or stratified medicine is understood as a medical approach in which patients are stratified based on their disease subtype, risk, prognosis, or treatment response using specialized diagnostic tests. The key idea is to base medical decisions on individual patient characteristics, including molecular and behavioral biomarkers, rather than on population averages. Personalized medicine is deeply connected to and dependent on data science, specifically machine learning (often named Artificial Intelligence in the mainstream media). While during recent years there has been a lot of enthusiasm about the potential of ‘big data’ and machine learning-based solutions, there exist only few examples that impact current clinical practice. The lack of impact on clinical practice can largely be attributed to insufficient performance of predictive models, difficulties to interpret complex model predictions, and lack of validation via prospective clinical trials that demonstrate a clear benefit compared to the standard of care. In this paper, we review the potential of state-of-the-art data science approaches for personalized medicine, discuss open challenges, and highlight directions that may help to overcome them in the future.ConclusionsThere is a need for an interdisciplinary effort, including data scientists, physicians, patient advocates, regulatory agencies, and health insurance organizations. Partially unrealistic expectations and concerns about data science-based solutions need to be better managed. In parallel, computational methods must advance more to provide direct benefit to clinical practice.

SummaryPluripotent stem cells can self-renew in culture and differentiate along all somatic lineages in vivo. While much is known about the molecular basis of pluripotency, the mechanisms of differentiation remain unclear. Here, we profile individual mouse embryonic stem cells as they progress along the neuronal lineage. We observe that cells pass from the pluripotent state to the neuronal state via an intermediate epiblast-like state. However, analysis of the rate at which cells enter and exit these observed cell states using a hidden Markov model indicates the presence of a chain of unobserved molecular states that each cell transits through stochastically in sequence. This chain of hidden states allows individual cells to record their position on the differentiation trajectory, thereby encoding a simple form of cellular memory. We suggest a statistical mechanics interpretation of these results that distinguishes between functionally distinct cellular “macrostates” and functionally similar molecular “microstates” and propose a model of stem cell differentiation as a non-Markov stochastic process.

A number of key regulators of mouse embryonic stem (ES) cell identity, including the transcription factor Nanog, show strong expression fluctuations at the single cell level. The molecular basis for these fluctuations is unknown. Here we used a genetic complementation strategy to investigate expression changes during transient periods of Nanog downregulation. Employing an integrated approach, that includes high-throughput single cell transcriptional profiling and mathematical modelling, we found that early molecular changes subsequent to Nanog loss are stochastic and reversible. However, analysis also revealed that Nanog loss severely compromises the self-sustaining feedback structure of the ES cell regulatory network. Consequently, these nascent changes soon become consolidated to committed fate decisions in the prolonged absence of Nanog. Consistent with this, we found that exogenous regulation of Nanog-dependent feedback control mechanisms produced more a homogeneous ES cell population. Taken together our results indicate that Nanog-dependent feedback loops have a role in controlling both ES cell fate decisions and population variability.