Causality in digital medicine

Ben Glocker (an expert in machine learning for medical imaging, Imperial College London), Mirco Musolesi (a data science and digital health expert, University College London), Jonathan Richens (an expert in diagnostic machine learning models, Babylon Health) and Caroline Uhler (a computational biology expert, MIT) talked to Nature Communications about their research interests in causality inference and how this can provide a robust framework for digital medicine studies and their implementation, across different fields of application.

Granger-Causality Inference of the Existence of Unobserved Important Components in Network Analysis

Detecting causal interrelationships in multivariate systems, in terms of the Granger-causality concept, is of major interest for applications in many fields. Analyzing all the relevant components of a system is almost impossible, which contrasts with the concept of Granger causality. Not observing some components might, in turn, lead to misleading results, particularly if the missing components are the most influential and important in the system under investigation. In networks, the importance of a node depends on the number of nodes connected to this node. The degree of centrality is the most commonly used measure to identify important nodes in networks. There are two kinds of degree centrality, which are in-degree and out-degree. This manuscrpt is concerned with finding the highest out-degree among nodes to identify the most influential nodes. Inferring the existence of unobserved important components is critical in many multivariate interacting systems. The implications of such a situation are discussed in the Granger-causality framework. To this end, two of the most recent Granger-causality techniques, renormalized partial directed coherence and directed partial correlation, were employed. They were then compared in terms of their performance according to the extent to which they can infer the existence of unobserved important components. Sub-network analysis was conducted to aid these two techniques in inferring the existence of unobserved important components, which is evidenced in the results. By comparing the results of the two conducted techniques, it can be asserted that renormalized partial coherence outperforms directed partial correlation in the inference of existing unobserved important components that have not been included in the analysis. This measure of Granger causality and sub-network analysis emphasizes their ubiquitous successful applicability in such cases of the existence of hidden unobserved important components.

Gerea: Prediction Of Gene Expression Regulators From Transcriptome Profiling Data To Transition Networks

Background: Mammalian genes are regulated at the transcriptional and post-transcriptional levels. These mechanisms may involve the direct promotion or inhibition of transcription via a regulator or post-transcriptional regulation through factors such as micro (mi)RNAs. Objective: This study aimed to construct gene regulation relationships modulated by causality inference-based miRNA-(transition factor)-(target gene) networks and analyze gene expression data to identify gene expression regulators. Methods: Mouse gene expression regulation relationships were manually curated from literature using a text mining method which was then employed to generate miRNA-(transition factor)-(target gene) networks. An algorithm was then introduced to identify gene expression regulators from transcriptome profiling data by applying enrichment analysis to these networks. Results: A total of 22,271 mouse gene expression regulation relationships were curated for 4,018 genes and 242 miRNAs. GEREA software was developed to perform the integrated analyses. We applied the algorithm to transcriptome data for synthetic miR-155 oligo-treated mouse CD4+ T-cells and confirmed that miR-155 is an important network regulator. The software was also tested on publicly available transcriptional profiling data for Salmonella infection, resulting in the identification of miR-125b as an important regulator. Conclusion: The causality inference-based miRNA-(transition factor)-(target gene) networks serve as a novel resource for gene expression regulation research, and GEREA is an effective and useful adjunct to the currently available methods. The regulatory networks and the algorithm implemented in the GEREA software package are available under a free academic license at website : http://www.thua45.cn/gerea.

Pervasive false brain connectivity from electrophysiological signals

1.AbstractSignals of brain electric neuronal activity, either invasively measured or non-invasively estimated, are commonly used for connectivity inference. One popular methodology assumes that the neural dynamics follow a multivariate autoregression, where the autoregressive coefficients represent the couplings among regions. If observation noise is present and ignored, as is common in practice, the estimated couplings are biased, affecting all forms of Granger-causality inference, both in time and in frequency domains. Significant nonsense coupling, i.e., nonsense connectivity, can appear when in reality there is none, since there is always observation noise in two possible forms: measurement noise, and activity from other brain regions due to volume conduction and low spatial resolution. This problem is critical, and is currently not being addressed, calling into question the validity of many Granger-causality reports in the literature. An estimation method that accounts for noise is based on an overdetermined system of high-order multivariate Yule-Walker equations, which give reduced variance estimators for the coupling coefficients of the unobserved signals. Simulation-based comparisons to other published methods are presented, demonstrating its adequate performance. In addition, simulation results are presented for a zero connectivity case with noisy observations, where the new method correctly reports no connectivity while classical analyses (as found in most software packages) report nonsense connectivity. For the sake of reproducible research, the supplementary material includes, in human readable format, all the time series data used here.

Dense time-course gene expression profiling of the Drosophila melanogaster innate immune response

ABSTRACTImmune responses need to be initiated rapidly, and maintained as needed, to prevent establishment and growth of infections. Still, immune genes differ in both initiation kinetics and shutdown dynamics. Here, we performed an RNA-seq time course on D. melanogaster with 20 time points post-LPS injection. A combination of methods, including spline fitting, cluster analysis, and Granger Causality inference, allowed detailed dissection of expression profiles and functional annotation of genes through guilt-by-association. We identified antimicrobial peptides as immediate-early response genes with a sustained up-regulation up to five days after stimulation, and genes in the IM family as having early and transient responses. We further observed a strong trade-off with metabolic genes, which strikingly recovered to pre-infection levels before the immune response was fully resolved. This high-dimensional dataset enables the comprehensive study of immune response dynamics through the parallel application of multiple temporal data analysis methods.