Check out our highlights from the PLOS Computational Biology August issue:
A systems approach reveals distinct metabolic strategies among the NCI-60
cancer cell lines
Altered metabolism is characteristic of many human diseases, including cancer, but disease progression and treatment efficacy vary between patients. Hence we need personalized approaches to define metabolic disease phenotypes and enable us to unravel the underlying disease mechanisms and to treat individuals more efficiently. Computational modelling increasingly supports the analysis of disease mechanisms and complex data sets, and the interpretation of extracellular metabolomic data sets is particularly promising since this data type is proximal to the actual metabolic phenotype altered in human diseases. Moreover, it might enable the direct interpretation of disease states from serum samples in the future. In this study Ines Thiele and colleagues take the first step towards this goal, by generating a large set of cancer metabolic models from extracellular metabolomic data and computationally stratifying the models based on their metabolic characteristics into different phenotype groups. Melanoma emerged as an interesting example of how their approach can provide insights into the intracellular metabolism from extracellular measurements. Taken together, this work paves the way to generate condition-specific models from extracellular metabolomic data and demonstrates the many ways by which distinct phenotypes can be stratified and phenotype-specific intervention targets can be predicted.
Neuron’s eye view: Inferring features of complex stimuli from neural responses
Many neuroscience experiments begin with a set of deliberately simplified stimuli designed to vary only along a small set of variables. Yet many phenomena of interest—natural movies, objects—are not easily parameterized by a small number of dimensions. Here, John M. Pearson and colleagues develop a novel Bayesian model for clustering stimuli based solely on neural responses, allowing us to discover which latent features of complex stimuli actually drive neural activity. They demonstrate that this model allows us to recover key features of neural responses in a pair of well-studied paradigms.
Multidomain analyses of a longitudinal human microbiome intestinal cleanout perturbation experiment
Complex dynamics of microbial communities underlie their essential roles in health and disease. To maintain or restore healthy states, Susan P. Holmes and colleagues set out to better understand the nature and basis of stability in the gut microbiota, under normal and perturbed conditions. Stability, resilience, and response to perturbation are central topics in community ecology. Extreme perturbations such as near-complete loss of biomass from a system can reveal factors that influence community structure. Recognizing the return to baseline diversity and abundances of biomarkers in community-wide recovery after a disturbance can foster an understanding of the basic pillars of resilience that contribute to human health. The authors designed a densely sampled longitudinal experiment in human volunteers using transient non-inflammatory diarrhea as the perturbation. In order to uncover the essential players in the recovery process, they tailored new advances in ribosomal sequence variant detection and sparse multidomain analytics that incorporate phylogenetic structure. They show sparse meaningful multidimensional projections that exhibit the essential features in resilient recovery. This work shows how a carefully designed longitudinal study combining denoised ribosomal RNA sequence variants and metagenomic data can inform the taxa and processes involved in the recovery from loss of large proportions of intestinal biomass.