Towards perturbation prediction of biological networks using deep learning

Graph representation learning in biomedicine and healthcare

Networks-or graphs-are universal descriptors of systems of interacting elements. In biomedicine and healthcare, they can represent, for example, molecular interactions, signalling pathways, disease co-morbidities or healthcare systems. In this Perspective, we posit that representation learning can realize principles of network medicine, discuss successes and current limitations of the use of representation learning on graphs in biomedicine and healthcare, and outline algorithmic strategies that leverage the topology of graphs to embed them into compact vectorial spaces. We argue that graph representation learning will keep pushing forward machine learning for biomedicine and healthcare applications, including the identification of genetic variants underlying complex traits, the disentanglement of single-cell behaviours and their effects on health, the assistance of patients in diagnosis and treatment, and the development of safe and effective medicines.

A graph convolutional network for predicting COVID-19 dynamics in 190 regions/countries

Introduction:

Coronavirus disease (COVID-19) rapidly spread from Wuhan, China to other parts of China and other regions/countries around the world, resulting in a pandemic due to large populations moving through the massive transport hubs connecting all regions of China via railways and a major international airport. COVID-19 will remain a threat until safe and effective vaccines and antiviral drugs have been developed, distributed, and administered on a global scale. Thus, there is urgent need to establish effective implementation of preemptive non-pharmaceutical interventions for appropriate prevention and control strategies, and predicting future COVID-19 cases is required to monitor and control the issue.

Methods

This study attempts to utilize a three-layer graph convolutional network (GCN) model to predict future COVID-19 cases in 190 regions and countries using COVID-19 case data, commercial flight route data, and digital maps of public transportation in terms of transnational human mobility. We compared the performance of the proposed GCN model to a multilayer perceptron (MLP) model on a dataset of COVID-19 cases (excluding the graph representation). The prediction performance of the models was evaluated using the mean squared error.

Results

Our results demonstrate that the proposed GCN model can achieve better graph utilization and performance compared to the baseline in terms of both prediction accuracy and stability.

Discussion

The proposed GCN model is a useful means to predict COVID-19 cases at regional and national levels. Such predictions can be used to facilitate public health solutions in public health responses to the COVID-19 pandemic using deep learning and data pooling. In addition, the proposed GCN model may help public health policymakers in decision making in terms of epidemic prevention and control strategies.

scTenifoldKnk: An efficient virtual knockout tool for gene function predictions via single-cell gene regulatory network perturbation

Gene knockout (KO) experiments are a proven, powerful approach for studying gene function. However, systematic KO experiments targeting a large number of genes are usually prohibitive due to the limit of experimental and animal resources. Here, we present scTenifoldKnk, an efficient virtual KO tool that enables systematic KO investigation of gene function using data from single-cell RNA sequencing (scRNA-seq). In scTenifoldKnk analysis, a gene regulatory network (GRN) is first constructed from scRNA-seq data of wild-type samples, and a target gene is then virtually deleted from the constructed GRN. Manifold alignment is used to align the resulting reduced GRN to the original GRN to identify differentially regulated genes, which are used to infer target gene functions in analyzed cells. We demonstrate that the scTenifoldKnk-based virtual KO analysis recapitulates the main findings of real-animal KO experiments and recovers the expected functions of genes in relevant cell types.

Machine learning methods trained on simple models can predict critical transitions in complex natural systems

Forecasting sudden changes in complex systems is a critical but challenging task, with previously developed methods varying widely in their reliability. Here we develop a novel detection method, using simple theoretical models to train a deep neural network to detect critical transitions-the Early Warning Signal Network (EWSNet). We then demonstrate that this network, trained on simulated data, can reliably predict observed real-world transitions in systems ranging from rapid climatic change to the collapse of ecological populations. Importantly, our model appears to capture latent properties in time series missed by previous warning signals approaches, allowing us to not only detect if a transition is approaching, but critically whether the collapse will be catastrophic or non-catastrophic. These novel properties mean EWSNet has the potential to serve as an indicator of transitions across a broad spectrum of complex systems, without requiring information on the structure of the system being monitored. Our work highlights the practicality of deep learning for addressing further questions pertaining to ecosystem collapse and has much broader management implications.

Group Behavior and Emergence of Cancer Drug Resistance

Drug resistance is a major impediment in cancer. Although it is generally thought that acquired drug resistance is due to genetic mutations, emerging evidence indicates that nongenetic mechanisms also play an important role. Resistance emerges through a complex interplay of clonal groups within a heterogeneous tumor and the surrounding microenvironment. Traits such as phenotypic plasticity, intercellular communication, and adaptive stress response, act in concert to ensure survival of intermediate reversible phenotypes, until permanent, resistant clones can emerge. Understanding the role of group behavior, and the underlying nongenetic mechanisms, can lead to more efficacious treatment designs and minimize or delay emergence of resistance.

Systems biology approach suggests new miRNAs as phenotypic stability factors in the epithelial-mesenchymal transition

The epithelial-mesenchymal transition (EMT) is a cellular programme on which epithelial cells undergo a phenotypic transition to mesenchymal ones acquiring metastatic properties such as mobility and invasion. TGF-β signalling can promote the EMT process. However, the dynamics of the concentration response of TGF-β-induced EMT is not well explained. In this work, we propose a logical model of TGF-β dose dependence of EMT in MCF10A breast cells. The model outcomes agree with experimentally observed phenotypes for the wild-type and perturbed/mutated cases. As important findings of the model, it predicts the coexistence of more than one hybrid state and that the circuit between TWIST1 and miR-129 is involved in their stabilization. Thus, miR-129 should be considered as a phenotypic stability factor. The circuit TWIST1/miR-129 associates with ZEB1-mediated circuits involving miRNAs 200, 1199, 340, and the protein GRHL2 to stabilize the hybrid state. Additionally, we found a possible new autocrine mechanism composed of the circuit involving TGF-β, miR-200, and SNAIL1 that contributes to the stabilization of the mesenchymal state. Therefore, our work can extend our comprehension of TGF-β-induced EMT in MCF10A cells to potentially improve the strategies for breast cancer treatment.

Network Approaches for Dissecting the Immune System

The immune system is a complex biological network composed of hierarchically organized genes, proteins, and cellular components that combat external pathogens and monitor the onset of internal disease. To meet and ultimately defeat these challenges, the immune system orchestrates an exquisitely complex interplay of numerous cells, often with highly specialized functions, in a tissue-specific manner. One of the major methodologies of systems immunology is to measure quantitatively the components and interaction levels in the immunologic networks to construct a computational network and predict the response of the components to perturbations. The recent advances in high-throughput sequencing techniques have provided us with a powerful approach to dissecting the complexity of the immune system. Here we summarize the latest progress in integrating omics data and network approaches to construct networks and to infer the underlying signaling and transcriptional landscape, as well as cell-cell communication, in the immune system, with a focus on hematopoiesis, adaptive immunity, and tumor immunology. Understanding the network regulation of immune cells has provided new insights into immune homeostasis and disease, with important therapeutic implications for inflammation, cancer, and other immune-mediated disorders.

BioModels-15 years of sharing computational models in life science

Computational modelling has become increasingly common in life science research. To provide a platform to support universal sharing, easy accessibility and model reproducibility, BioModels (https://www.ebi.ac.uk/biomodels/), a repository for mathematical models, was established in 2005. The current BioModels platform allows submission of models encoded in diverse modelling formats, including SBML, CellML, PharmML, COMBINE archive, MATLAB, Mathematica, R, Python or C++. The models submitted to BioModels are curated to verify the computational representation of the biological process and the reproducibility of the simulation results in the reference publication. The curation also involves encoding models in standard formats and annotation with controlled vocabularies following MIRIAM (minimal information required in the annotation of biochemical models) guidelines. BioModels now accepts large-scale submission of auto-generated computational models. With gradual growth in content over 15 years, BioModels currently hosts about 2000 models from the published literature. With about 800 curated models, BioModels has become the world’s largest repository of curated models and emerged as the third most used data resource after PubMed and Google Scholar among the scientists who use modelling in their research. Thus, BioModels benefits modellers by providing access to reliable and semantically enriched curated models in standard formats that are easy to share, reproduce and reuse.