Gene networks

Graph Node Classification to Predict Autism Risk in Genes

This study explores the genetic risk associations with autism spectrum disorder (ASD) using graph neural networks (GNNs), leveraging the Sfari dataset and protein interaction network (PIN) data. We built a gene network with genes as nodes, chromosome band location as node features, and gene interactions as edges. Graph models were employed to classify the autism risk associated with newly introduced genes (test set). Three classification tasks were undertaken to test the ability of our models: binary risk association, multi-class risk association, and syndromic gene association. We tested graph convolutional networks, Graph Sage, graph transformer, and Multi-Layer Perceptron (Baseline) architectures on this problem. The Graph Sage model consistently outperformed the other models, showcasing its utility in classifying ASD-related genes. Our ablation studies show that the chromosome band location and protein interactions contain useful information for this problem. The models achieved 85.80% accuracy on the binary risk classification, 81.68% accuracy on the multi-class risk classification, and 90.22% on the syndromic classification.

Urinary exosomal hsa_circ_0001250 as a novel diagnostic biomarker of idiopathic membranous nephropathy

AIMS

Idiopathic membranous nephropathy (IMN) is a common cause of adult nephrotic syndrome. Currently, the diagnosis of IMN mainly depends on renal biopsy, which is invasive. What's more, markers already known for the clinical diagnosis of IMN are not sensitive enough. The present study aims to investigate the profiling of urinary exosomal circular RNAs (circRNAs) of IMN, and to look for a potential biomarker for diagnosis of IMN.

METHODS

Urine exosomes were collected from patients with IMN and idiopathic nephrotic syndrome (INS), as well as healthy controls (HCs) by ultracentrifuge. A pairwise comparison between 5 IMN and 5 HC was performed by high-throughput sequencing. Enrichment analysis were performed to explore the potential functions of differentially expressed circRNAs in IMN. Among three differentially expressed circRNAs which may be involved in signaling pathways of pathogenesis of IMN and matched conserved mouse circRNAs, hsa_circ_0001250 was selected as the target circRNA after quantitative polymerase chain reaction among 23 IMN, 19 INS and 23HC. Sanger sequencing and RNase R digestion assay were performed to validated the ring-structure and sequence of hsa_circ_0001250. ROC (Receiver Operating Characteristic) curve correlation analysis was used to further validate the potential utility of hsa_circ_0001250 as a diagnostic biomarker of IMN. A circRNA-miRNA-mRNA network was constructed to reflect the relationship between hsa_circ_0001250 and its target miRNAs and mRNAs.

RESULTS

766 up-regulated and 283 down-regulated circRNAs were identified in IMN patients. Kyoto Encyclopedia of Genes and Genomes pathway analysis revealed signaling pathways of pathogenesis of IMN which the different expressed circRNAs may participate in. The ring-structure and the sequence of hsa_circ_0001250 were confirmed, the expression of hsa_circ_0001250 was validated significantly increased in IMN, relevant with high level of proteinuria. A circRNA-miRNA-mRNA network reflected that hsa_circ_0001250 may play a role in the pathogenesis of IMN by target hsa-miR-639 and hsa-miR-4449.

CONCLUSION

We revealed the expression and functional profile of differentially expressed urinary exosomal circRNAs of IMN patients. Urinary exosomal hsa_circ_0001250 was tested as a potential biomarker of IMN and a predicted circRNA-miRNA-mRNA network was constructed.

A compendium of human genes regulating feeding behavior and body weight, its functional characterization and identification of GWAS genes involved in brain-specific PPI network

Background

Obesity is heritable. It predisposes to many diseases. The objectives of this study were to create a compendium of genes relevant to feeding behavior (FB) and/or body weight (BW) regulation; to construct and to analyze networks formed by associations between genes/proteins; and to identify the most significant genes, biological processes/pathways, and tissues/organs involved in BW regulation.

Results

The compendium of genes controlling FB or BW includes 578 human genes. Candidate genes were identified from various sources, including previously published original research and review articles, GWAS meta-analyses, and OMIM (Online Mendelian Inheritance in Man). All genes were ranked according to knowledge about their biological role in body weight regulation and classified according to expression patterns or functional characteristics. Substantial and overrepresented numbers of genes from the compendium encoded cell surface receptors, signaling molecules (hormones, neuropeptides, cytokines), transcription factors, signal transduction proteins, cilium and BBSome components, and lipid binding proteins or were present in the brain-specific list of tissue-enriched genes identified with TSEA tool. We identified 27 pathways from KEGG, REACTOME and BIOCARTA whose genes were overrepresented in the compendium. Networks formed by physical interactions or homological relationships between proteins or interactions between proteins involved in biochemical/signaling pathways were reconstructed and analyzed. Subnetworks and clusters identified by the MCODE tool included genes/proteins associated with cilium morphogenesis, signal transduction proteins (particularly, G protein–coupled receptors, kinases or proteins involved in response to insulin stimulus) and transcription regulation (particularly nuclear receptors). We ranked GWAS genes according to the number of neighbors in three networks and revealed 22 GWAS genes involved in the brain-specific PPI network. On the base of the most reliable PPIs functioning in the brain tissue, new regulatory schemes interpreting relevance to BW regulation are proposed for three GWAS genes (ETV5, LRP1B, and NDUFS3).

Conclusions

A compendium comprising 578 human genes controlling FB or BW was designed, and the most significant functional groups of genes, biological processes/pathways, and tissues/organs involved in BW regulation were revealed. We ranked genes from the GWAS meta-analysis set according to the number and quality of associations in the networks and then according to their involvement in the brain-specific PPI network and proposed new regulatory schemes involving three GWAS genes (ETV5, LRP1B, and NDUFS3) in BW regulation. The compendium is expected to be useful for pathology risk estimation and for design of new pharmacological approaches in the treatment of human obesity.

Electronic supplementary material

The online version of this article (doi:10.1186/s12863-016-0466-2) contains supplementary material, which is available to authorized users.

Novel tuberculosis susceptibility candidate genes revealed by the reconstruction and analysis of associative networks

Tuberculosis (TB) is a common infectious disease caused by M. tuberculosis. The risk of the disease is dependent on complex interactions between host genetics and environmental factors. Accumulated genomic data, along with novel methodological approaches such as associative networks, facilitate studies into the inherited basis of TB. In the current study, we carried out the reconstruction and analysis of an associative network representing molecular interactions between proteins and genes associated with TB. The network predominantly comprises of well-studied key proteins and genes which are able to govern the immune response against M. tuberculosis. However, this approach also allowed us to reveal 12 proteins encoded by genes, the polymorphisms of which have never been studied in relation to M. tuberculosis infection. These proteins include surface antigens (CD4, CD69, CD79, CD80, MUC16) and other important components of the immune response, inflammation, pathogen recognition, cell migration and activation (HCST, ADA, CP, SPP1, CXCR4, AGER, PACRG). Thus, the associative network approach enables the discovery of new candidate genes for TB susceptibility.

ON THE CHAOS IN GENE NETWORKS

The methods for constructing "chaotic" nonlinear systems of differential equations modeling gene networks of arbitrary structure and dimensionality with various types of symmetry are considered. It has been shown that an increase in modality of the functions describing the control of gene expression efficiency allows for a decrease in the dimensionality of these systems with retention of their chaotic dynamics. Three-dimensional "chaotic" cyclic systems are considered. Symmetrical and asymmetrical attractors with "narrow" chaos having a Moebius-like structure have been detected in such systems. As has been demonstrated, a complete symmetry of the systems with respect to permutation of variables does not prevent the emergence of their chaotic dynamics.

Netscan: a procedure for generating reaction networks by size

In this paper, we describe an algorithm which can be used to generate the collection of networks, in order of increasing size, that are compatible with a list of chemical reactions and that satisfy a constraint. Our algorithm has been encoded and the software, called Netscan, can be freely downloaded from ftp://ftp.stat.ubc.ca/pub/riffraff/Netscanfiles, along with a manual, for general scientific use. Our algorithm may require pre-processing to ensure that the quantities it acts on are physically relevant and because it outputs sets of reactions, which we call canonical networks, that must be elaborated into full networks.