Multiple imputations applied to the DREAM3 phosphoproteomics challenge: a winning strategy

Virtual Populations for Quantitative Systems Pharmacology Models

Quantitative systems pharmacology (QSP) places an emphasis on dynamic systems modeling, incorporating considerations from systems biology modeling and pharmacodynamics. The goal of QSP is often to quantitatively predict the effects of clinical therapeutics, their combinations, and their doses on clinical biomarkers and endpoints. In order to achieve this goal, strategies for incorporating clinical data into model calibration are critical. Virtual population (VPop) approaches facilitate model calibration while faced with challenges encountered in QSP model application, including modeling a breadth of clinical therapies, biomarkers, endpoints, utilizing data of varying structure and source, capturing observed clinical variability, and simulating with models that may require more substantial computational time and resources than often found in pharmacometrics applications. VPops are frequently developed in a process that may involve parameterization of isolated pathway models, integration into a larger QSP model, incorporation of clinical data, calibration, and quantitative validation that the model with the accompanying, calibrated VPop is suitable to address the intended question or help with the intended decision. Here, we introduce previous strategies for developing VPops in the context of a variety of therapeutic and safety areas: metabolic disorders, drug-induced liver injury, autoimmune diseases, and cancer. We introduce methodological considerations, prior work for sensitivity analysis and VPop algorithm design, and potential areas for future advancement. Finally, we give a more detailed application example of a VPop calibration algorithm that illustrates recent progress and many of the methodological considerations. In conclusion, although methodologies have varied, VPop strategies have been successfully applied to give valid clinical insights and predictions with the assistance of carefully defined and designed calibration and validation strategies. While a uniform VPop approach for all potential QSP applications may be challenging given the heterogeneity in use considerations, we anticipate continued innovation will help to drive VPop application for more challenging cases of greater scale while developing new rigorous methodologies and metrics.

A simple prognostic index based on admission vital signs data among patients with sepsis in a resource-limited setting

Introduction

In sub-Saharan Africa, vital signs are a feasible option for monitoring critically ill patients. We assessed how admission vital signs data predict in-hospital mortality among patients with sepsis. In particular, we assessed whether vital signs data can be incorporated into a prognostic index with reduced segmentation in the values of included variables.

Methods

Subjects were patients with sepsis hospitalized in Uganda, who participated in two cohort studies. Using restricted cubic splines of admission vital signs data, we predicted probability of in-hospital death in the development cohort and used this information to construct a simple prognostic index. We assessed the performance of the index in a validation cohort and compared its performance to that of the Modified Early Warning Score (MEWS).

Results

We included 317 patients (167 in the development cohort and 150 in the validation cohort). Based on how vital signs predicted mortality, we created a prognostic index giving a score of 1 for: respiratory rates ≥30 cycles/minute; pulse rates ≥100 beats/minute; mean arterial pressures ≥110/<70 mmHg; temperatures ≥38.6/<35.6°C; and presence of altered mental state defined as Glasgow coma score ≤14; 0 for all other values. The proposed index (maximum score = 5) predicted mortality comparably to MEWS. Patients scoring ≥3 on the index were 3.4-fold (95% confidence interval (CI) 1.6 to 7.3, P = 0.001) and 2.3-fold (95% CI 1.1 to 4.7, P = 0.031) as likely to die in hospital as those scoring 0 to 2 in the development and validation cohorts respectively; those scoring ≥5 on MEWS were 2.5-fold (95% CI 1.2 to 5.3, P = 0.017) and 1.8-fold (95% CI 0.74 to 4.2, P = 0.204) as likely to die as those scoring 0 to 4 in the development and validation cohorts respectively.

Conclusion

Among patients with sepsis, a prognostic index incorporating admission vital signs data with reduced segmentation in the values of included variables adequately predicted mortality. Such an index may be more easily implemented when triaging acutely-ill patients. Future studies using a similar approach may develop indexes that can be used to monitor treatment among acutely-ill patients, especially in resource-limited settings.

Data- and knowledge-based modeling of gene regulatory networks: an update

Gene regulatory network inference is a systems biology approach which predicts interactions between genes with the help of high-throughput data. In this review, we present current and updated network inference methods focusing on novel techniques for data acquisition, network inference assessment, network inference for interacting species and the integration of prior knowledge. After the advance of Next-Generation-Sequencing of cDNAs derived from RNA samples (RNA-Seq) we discuss in detail its application to network inference. Furthermore, we present progress for large-scale or even full-genomic network inference as well as for small-scale condensed network inference and review advances in the evaluation of network inference methods by crowdsourcing. Finally, we reflect the current availability of data and prior knowledge sources and give an outlook for the inference of gene regulatory networks that reflect interacting species, in particular pathogen-host interactions.

Identifying biological network structure, predicting network behavior, and classifying network state with High Dimensional Model Representation (HDMR)

This work presents an adapted Random Sampling - High Dimensional Model Representation (RS-HDMR) algorithm for synergistically addressing three key problems in network biology: (1) identifying the structure of biological networks from multivariate data, (2) predicting network response under previously unsampled conditions, and (3) inferring experimental perturbations based on the observed network state. RS-HDMR is a multivariate regression method that decomposes network interactions into a hierarchy of non-linear component functions. Sensitivity analysis based on these functions provides a clear physical and statistical interpretation of the underlying network structure. The advantages of RS-HDMR include efficient extraction of nonlinear and cooperative network relationships without resorting to discretization, prediction of network behavior without mechanistic modeling, robustness to data noise, and favorable scalability of the sampling requirement with respect to network size. As a proof-of-principle study, RS-HDMR was applied to experimental data measuring the single-cell response of a protein-protein signaling network to various experimental perturbations. A comparison to network structure identified in the literature and through other inference methods, including Bayesian and mutual-information based algorithms, suggests that RS-HDMR can successfully reveal a network structure with a low false positive rate while still capturing non-linear and cooperative interactions. RS-HDMR identified several higher-order network interactions that correspond to known feedback regulations among multiple network species and that were unidentified by other network inference methods. Furthermore, RS-HDMR has a better ability to predict network response under unsampled conditions in this application than the best statistical inference algorithm presented in the recent DREAM3 signaling-prediction competition. RS-HDMR can discern and predict differences in network state that arise from sources ranging from intrinsic cell-cell variability to altered experimental conditions, such as when drug perturbations are introduced. This ability ultimately allows RS-HDMR to accurately classify the experimental conditions of a given sample based on its observed network state.

Towards a rigorous assessment of systems biology models: the DREAM3 challenges

Background

Systems biology has embraced computational modeling in response to the quantitative nature and increasing scale of contemporary data sets. The onslaught of data is accelerating as molecular profiling technology evolves. The Dialogue for Reverse Engineering Assessments and Methods (DREAM) is a community effort to catalyze discussion about the design, application, and assessment of systems biology models through annual reverse-engineering challenges.

Methodology and Principal Findings

We describe our assessments of the four challenges associated with the third DREAM conference which came to be known as the DREAM3 challenges: signaling cascade identification, signaling response prediction, gene expression prediction, and the DREAM3 in silico network challenge. The challenges, based on anonymized data sets, tested participants in network inference and prediction of measurements. Forty teams submitted 413 predicted networks and measurement test sets. Overall, a handful of best-performer teams were identified, while a majority of teams made predictions that were equivalent to random. Counterintuitively, combining the predictions of multiple teams (including the weaker teams) can in some cases improve predictive power beyond that of any single method.

Conclusions

DREAM provides valuable feedback to practitioners of systems biology modeling. Lessons learned from the predictions of the community provide much-needed context for interpreting claims of efficacy of algorithms described in the scientific literature.