HIPPIE: Integrating protein interaction networks with experiment based quality scores

Prediction and Evaluation of Coronavirus and Human Protein-Protein Interactions Integrating Five Different Computational Methods

The high lethality and infectiousness of coronaviruses, particularly SARS-Cov-2, pose a significant threat to human society. Understanding coronaviruses, especially the interactions between these viruses and humans, is crucial for mitigating the coronavirus pandemic. In this study, we conducted a comprehensive comparison and evaluation of five prevalent computational methods: interolog mapping, domain-domain interaction methodology, domain-motif interaction methodology, structure-based approaches, and machine learning techniques. These methods were assessed using unbiased datasets that include C1, C2h, C2v, and C3 test sets. Ultimately, we integrated these five methodologies into a unified model for predicting protein-protein interactions (PPIs) between coronaviruses and human proteins. Our final model demonstrates relatively better performance, particularly with the C2v and C3 test sets, which are frequently used datasets in practical applications. Based on this model, we further established a high-confidence PPI network between coronaviruses and humans, consisting of 18,012 interactions between 3843 human proteins and 129 coronavirus proteins. The reliability of our predictions was further validated through the current knowledge framework and network analysis. This study is anticipated to enhance mechanistic understanding of the coronavirus-human relationship a while facilitating the rediscovery of antiviral drug targets. The source codes and datasets are accessible at https://github.com/covhppilab/CoVHPPI.

A Map of the Lipid-Metabolite-Protein Network to Aid Multi-Omics Integration

The integration of multi-omics data offers transformative potential for elucidating complex molecular mechanisms underlying biological processes and diseases. In this study, we developed a lipid-metabolite-protein network that combines a protein-protein interaction network and enzymatic and genetic interactions of proteins with metabolites and lipids to provide a unified framework for multi-omics integration. Using hyperbolic embedding, the network visualizes connections across omics layers, accessible through a user-friendly Shiny R (version 1.10.0) software package. This framework ranks molecules across omics layers based on functional proximity, enabling intuitive exploration. Application in a cardiovascular disease (CVD) case study identified lipids and metabolites associated with CVD-related proteins. The analysis confirmed known associations, like cholesterol esters and sphingomyelin, and highlighted potential novel biomarkers, such as 4-imidazoleacetate and indoleacetaldehyde. Furthermore, we used the network to analyze empagliflozin's temporal effects on lipid metabolism. Functional enrichment analysis of proteins associated with lipid signatures revealed dynamic shifts in biological processes, with early effects impacting phospholipid metabolism and long-term effects affecting sphingolipid biosynthesis. Our framework offers a versatile tool for hypothesis generation, functional analysis, and biomarker discovery. By bridging molecular layers, this approach advances our understanding of disease mechanisms and therapeutic effects, with broad applications in computational biology and precision medicine.

Prediction of protein interactions with function in protein (de-)phosphorylation

Protein-protein interactions (PPIs) form a complex network called "interactome" that regulates many functions in the cell. In recent years, there is an increasing accumulation of evidence supporting the existence of a hyperbolic geometry underlying the network representation of complex systems such as the interactome. In particular, it has been shown that the embedding of the human Protein-Interaction Network (hPIN) in hyperbolic space (H2) captures biologically relevant information. Here we explore whether this mapping contains information that would allow us to predict the function of PPIs, more specifically interactions related to post-translational modification (PTM). We used a random forest algorithm to predict PTM-related directed PPIs, concretely, protein phosphorylation and dephosphorylation, based on hyperbolic properties and centrality measures of the hPIN mapped in H2. To evaluate the efficacy of our algorithm, we predicted PTM-related PPIs of ataxin-1, a protein which is responsible for Spinocerebellar Ataxia type 1 (SCA1). Proteomics analysis in a cellular model revealed that several of the predicted PTM-PPIs were indeed dysregulated in a SCA1-related disease network. A compact cluster composed of ataxin-1, its dysregulated PTM-PPIs and their common upstream regulators may represent critical interactions for disease pathology. Thus, our algorithm may infer phosphorylation activity on proteins through directed PPIs.

Advanced AI and ML frameworks for transforming drug discovery and optimization: With innovative insights in polypharmacology, drug repurposing, combination therapy and nanomedicine

Artificial Intelligence (AI) and Machine Learning (ML) are transforming drug discovery by overcoming traditional challenges like high costs, time-consuming, and frequent failures. AI-driven approaches streamline key phases, including target identification, lead optimization, de novo drug design, and drug repurposing. Frameworks such as deep neural networks (DNNs), convolutional neural networks (CNNs), and deep reinforcement learning (DRL) models have shown promise in identifying drug targets, optimizing delivery systems, and accelerating drug repurposing. Generative adversarial networks (GANs) and variational autoencoders (VAEs) aid de novo drug design by creating novel drug-like compounds with desired properties. Case studies, such as DDR1 kinase inhibitors designed using generative models and CDK20 inhibitors developed via structure-based methods, highlight AI's ability to produce highly specific therapeutics. Models like SNF-CVAE and DeepDR further advance drug repurposing by uncovering new therapeutic applications for existing drugs. Advanced ML algorithms enhance precision in predicting drug efficacy, toxicity, and ADME-Tox properties, reducing development costs and improving drug-target interactions. AI also supports polypharmacology by optimizing multi-target drug interactions and enhances combination therapy through predictions of drug synergies and antagonisms. In nanomedicine, AI models like CURATE.AI and the Hartung algorithm optimize personalized treatments by predicting toxicological risks and real-time dosing adjustments with high accuracy. Despite its potential, challenges like data quality, model interpretability, and ethical concerns must be addressed. High-quality datasets, transparent models, and unbiased algorithms are essential for reliable AI applications. As AI continues to evolve, it is poised to revolutionize drug discovery and personalized medicine, advancing therapeutic development and patient care.

Improving the identification of cancer driver modules using deep features learned from multi-omics data

Identifying the cancer driver modules or pathways is crucial to understanding the fundamental mechanisms of cancer occurrence and progression. The rapid abundance of cancer omics data provides unprecedented opportunities to study the driver modules in cancer, and many computational methods have been developed in recent years. However, most existing methods have limitations in considering different types of cancer omics data and cannot effectively learn informative omics features for integrated identification of driver modules. In this paper, we introduce a new integrated framework to accurately identify the cancer driver modules by integrating the protein-protein interaction network, transcriptional regulatory network, gene expression and mutation data in cancer. We first develop a series of methods to learn the deep features of functional connectivity between genes in each omics data and then construct an integrated gene functional coherence network. Furthermore, we present a two-step module mining method to efficiently identify the cancer driver modules from the integrated gene functional coherence network. Systematic experiments in three cancer types demonstrate that the proposed framework can obtain more significant driver modules than most existing methods, and some identified driver modules are associated with clinical survival phenotypes.

Docetaxel response in BRCA1,p53-deficient mammary tumor cells is affected by Huntingtin and BAP1

Taxanes are frequently used anticancer drugs known to kill tumor cells by inducing mitotic aberrations and segregation defects. A defining feature of specific cancers, notably triple-negative breast cancer (TNBC) and particularly those deficient in BRCA1, is chromosomal instability (CIN). Here, we focused on understanding the mechanisms of docetaxel-induced cytotoxicity, especially in the context of BRCA1-deficient TNBC. Using functional genetic screens in CIN+ cells, we identified genes that mediate docetaxel response and found an interaction between Huntingtin (HTT) and BRCA1-associated protein-1 (BAP1). We employed Brca1-/-;p53-/- mammary tumor cells, derived from genetically engineered mouse tumors that closely mimic the human disease, to investigate the role of these genes in CIN+ BRCA1-deficient cells. Specifically, we observed that loss of HTT sensitizes CIN+ BRCA1-deficient mammary tumor cells to docetaxel by shortening mitotic spindle poles and increasing spindle multipolarity. In contrast, BAP1 depletion protected cells against these spindle aberrations by restoring spindle length and enhancing mitotic clustering of the extra centrosomes. In conclusion, our findings shed light on the roles of HTT and BAP1 in controlling mitotic spindle multipolarity and centrosome clustering, specifically in the absence of BRCA1. This affects the response to microtubule-targeting agents and suggests that further studies of the interaction of these genes with the mitotic spindle may provide useful insights into how to target CIN+ cells, particularly in the challenging therapeutic landscape of BRCA1-deficient TNBC.

SENSE-PPI reconstructs interactomes within, across, and between species at the genome scale

Ab initio computational reconstructions of protein-protein interaction (PPI) networks will provide invaluable insights into cellular systems, enabling the discovery of novel molecular interactions and elucidating biological mechanisms within and between organisms. Leveraging the latest generation protein language models and recurrent neural networks, we present SENSE-PPI, a sequence-based deep learning model that efficiently reconstructs ab initio PPIs, distinguishing partners among tens of thousands of proteins and identifying specific interactions within functionally similar proteins. SENSE-PPI demonstrates high accuracy, limited training requirements, and versatility in cross-species predictions, even with non-model organisms and human-virus interactions. Its performance decreases for phylogenetically more distant model and non-model organisms, but signal alteration is very slow. In this regard, it demonstrates the important role of parameters in protein language models. SENSE-PPI is very fast and can test 10,000 proteins against themselves in a matter of hours, enabling the reconstruction of genome-wide proteomes.

Global Interactome Mapping Reveals Pro-tumorigenic Interactions of NF-κB in Breast Cancer

NF-κB pathway is involved in inflammation; however, recent data shows its role also in cancer development and progression, including metastasis. To understand the role of NF-κB interactome dynamics in cancer, we study the complexity of breast cancer interactome in luminal A breast cancer model and its rearrangement associated with NF-κB modulation. Liquid chromatography-mass spectrometry measurement of 160 size-exclusion chromatography fractions identifies 5460 protein groups. Seven thousand five hundred sixty eight interactions among these proteins have been reconstructed by PrInCE algorithm, of which 2564 have been validated in independent datasets. NF-κB modulation leads to rearrangement of protein complexes involved in NF-κB signaling and immune response, cell cycle regulation, and DNA replication. Central NF-κB transcription regulator RELA co-elutes with interactors of NF-κB activator PRMT5, and these complexes are confirmed by AlphaPulldown prediction. A complementary immunoprecipitation experiment recapitulates RELA interactions with other NF-κB factors, associating NF-κB inhibition with lower binding of NF-κB activators to RELA. This study describes a network of pro-tumorigenic protein interactions and their rearrangement upon NF-κB inhibition with potential therapeutic implications in tumors with high NF-κB activity.

A comprehensive review of protein-centric predictors for biomolecular interactions: from proteins to nucleic acids and beyond

Proteins interact with diverse ligands to perform a large number of biological functions, such as gene expression and signal transduction. Accurate identification of these protein-ligand interactions is crucial to the understanding of molecular mechanisms and the development of new drugs. However, traditional biological experiments are time-consuming and expensive. With the development of high-throughput technologies, an increasing amount of protein data is available. In the past decades, many computational methods have been developed to predict protein-ligand interactions. Here, we review a comprehensive set of over 160 protein-ligand interaction predictors, which cover protein-protein, protein-nucleic acid, protein-peptide and protein-other ligands (nucleotide, heme, ion) interactions. We have carried out a comprehensive analysis of the above four types of predictors from several significant perspectives, including their inputs, feature profiles, models, availability, etc. The current methods primarily rely on protein sequences, especially utilizing evolutionary information. The significant improvement in predictions is attributed to deep learning methods. Additionally, sequence-based pretrained models and structure-based approaches are emerging as new trends.

Active natural compounds perturb the melanoma risk-gene network

Cutaneous melanoma is an aggressive type of skin cancer with a complex genetic landscape caused by the malignant transformation of melanocytes. This study aimed at providing an in silico network model based on the systematic profiling of the melanoma-associated genes considering germline mutations, somatic mutations, and genome-wide association study signals accounting for a total of 232 unique melanoma risk genes. A protein-protein interaction network was constructed using the melanoma risk genes as seeds and evaluated to describe the functional landscape in which the melanoma genes operate within the cellular milieu. Not only were the majority of the melanoma risk genes able to interact with each other at the protein level within the core of the network, but this showed significant enrichment for genes whose expression is altered in human melanoma specimens. Functional annotation showed the melanoma risk network to be significantly associated with processes related to DNA metabolism and telomeres, DNA damage and repair, cellular ageing, and response to radiation. We further explored whether the melanoma risk network could be used as an in silico tool to predict the efficacy of anti-melanoma phytochemicals, that are considered active molecules with potentially less systemic toxicity than classical cytotoxic drugs. A significant portion of the melanoma risk network showed differential expression when SK-MEL-28 human melanoma cells were exposed to the phytochemicals harmine and berberine chloride. This reinforced our hypothesis that the network modeling approach not only provides an alternative way to identify molecular pathways relevant to disease but it may also represent an alternative screening approach to prioritize potentially active compounds.

A protein-proximity screen reveals Ebola virus co-opts the mRNA decapping complex through the scaffold protein EDC4

The interaction of host and Ebola virus (EBOV) proteins is required for establishing infection of the cell. To identify protein binding partners, a proximity-dependent protein interaction screen was performed for six EBOV proteins. Hits were computationally mapped onto a human protein-protein interactome and then annotated with viral proteins to reveal known and previously undescribed EBOV-host protein interactions and processes. Importantly, this approach efficiently arranged proteins into functional complexes associated with single viral proteins. Focused characterization of interactions between EBOV VP35 and the mRNA decapping complex demonstrated that VP35 binds the scaffold protein EDC4 through the C-terminal subdomain, with each protein found associated in EBOV-infected cells. Mechanistically, depletion of three components of the complex each similarly inhibited viral replication by reducing early viral RNA synthesis. Overall, we demonstrate successful identification of EBOV protein interaction with entire cellular machines, providing a deeper understanding of replication mechanism for therapeutic intervention.

Bioinformatics Prediction for Network-Based Integrative Multi-Omics Expression Data Analysis in Hirschsprung Disease

Hirschsprung's disease (HSCR) is a rare developmental disorder in which enteric ganglia are missing along a portion of the intestine. HSCR has a complex inheritance, with RET as the major disease-causing gene. However, the pathogenesis of HSCR is still not completely understood. Therefore, we applied a computational approach based on multi-omics network characterization and clustering analysis for HSCR-related gene/miRNA identification and biomarker discovery. Protein-protein interaction (PPI) and miRNA-target interaction (MTI) networks were analyzed by DPClusO and BiClusO, respectively, and finally, the biomarker potential of miRNAs was computationally screened by miRNA-BD. In this study, a total of 55 significant gene-disease modules were identified, allowing us to propose 178 new HSCR candidate genes and two biological pathways. Moreover, we identified 12 key miRNAs with biomarker potential among 137 predicted HSCR-associated miRNAs. Functional analysis of new candidates showed that enrichment terms related to gene ontology (GO) and pathways were associated with HSCR. In conclusion, this approach has allowed us to decipher new clues of the etiopathogenesis of HSCR, although molecular experiments are further needed for clinical validations.

Intranuclear inclusions of polyQ-expanded ATXN1 sequester RNA molecules

Spinocerebellar ataxia type 1 (SCA1) is an autosomal dominant neurodegenerative disease caused by a trinucleotide (CAG) repeat expansion in the ATXN1 gene. It is characterized by the presence of polyglutamine (polyQ) intranuclear inclusion bodies (IIBs) within affected neurons. In order to investigate the impact of polyQ IIBs in SCA1 pathogenesis, we generated a novel protein aggregation model by inducible overexpression of the mutant ATXN1(Q82) isoform in human neuroblastoma SH-SY5Y cells. Moreover, we developed a simple and reproducible protocol for the efficient isolation of insoluble IIBs. Biophysical characterization showed that polyQ IIBs are enriched in RNA molecules which were further identified by next-generation sequencing. Finally, a protein interaction network analysis indicated that sequestration of essential RNA transcripts within ATXN1(Q82) IIBs may affect the ribosome resulting in error-prone protein synthesis and global proteome instability. These findings provide novel insights into the molecular pathogenesis of SCA1, highlighting the role of polyQ IIBs and their impact on critical cellular processes.

An androgen receptor regulated gene score is associated with epithelial to mesenchymal transition features in triple negative breast cancers

BACKGROUND

Androgen receptor (AR) is considered a marker of better prognosis in hormone receptor positive breast cancers (BC), however, its role in triple negative breast cancer (TNBC) is controversial. This may be attributed to intrinsic molecular differences or scoring methods for AR positivity. We derived AR regulated gene score and examined its utility in BC subtypes.

METHODS

AR regulated genes were derived by applying a bioinformatic pipeline on publicly available microarray data sets of AR+ BC cell lines and gene score was calculated as average expression of six AR regulated genes. Tumors were divided into AR high and low based on gene score and associations with clinical parameters, circulating androgens, survival and epithelial to mesenchymal transition (EMT) markers were examined, further evaluated in invitro models and public datasets.

RESULTS

53% (133/249) tumors were classified as AR gene score high and were associated with significantly better clinical parameters, disease-free survival (86.13 vs 72.69 months, log rank p = 0.032) when compared to AR low tumors. 36% of TNBC (N = 66) were AR gene score high with higher expression of EMT markers (p = 0.024) and had high intratumoral levels of 5α-reductase, enzyme involved in intracrine androgen metabolism. In MDA-MB-453 treated with dihydrotestosterone, SLUG expression increased, E-cadherin decreased with increase in migration and these changes were reversed with bicalutamide. Similar results were obtained in public datasets.

CONCLUSION

Deciphering the role of AR in BC is difficult based on AR protein levels alone. Our results support the context dependent function of AR in driving better prognosis in ER positive tumors and EMT features in TNBC tumors.

From Pathogenesis to Therapeutics: A Review of 150 Years of Huntington's Disease Research

Huntington's disease (HD) is a debilitating neurodegenerative genetic disorder caused by an expanded polyglutamine-coding (CAG) trinucleotide repeat in the huntingtin (HTT) gene. HD behaves as a highly penetrant dominant disorder likely acting through a toxic gain of function by the mutant huntingtin protein. Widespread cellular degeneration of the medium spiny neurons of the caudate nucleus and putamen are responsible for the onset of symptomology that encompasses motor, cognitive, and behavioural abnormalities. Over the past 150 years of HD research since George Huntington published his description, a plethora of pathogenic mechanisms have been proposed with key themes including excitotoxicity, dopaminergic imbalance, mitochondrial dysfunction, metabolic defects, disruption of proteostasis, transcriptional dysregulation, and neuroinflammation. Despite the identification and characterisation of the causative gene and mutation and significant advances in our understanding of the cellular pathology in recent years, a disease-modifying intervention has not yet been clinically approved. This review includes an overview of Huntington's disease, from its genetic aetiology to clinical presentation and its pathogenic manifestation. An updated view of molecular mechanisms and the latest therapeutic developments will also be discussed.

Splicing Modulators Are Involved in Human Polyglutamine Diversification via Protein Complexes Shuttling between Nucleus and Cytoplasm

Length polymorphisms of polyglutamine (polyQs) in triplet-repeat-disease-causing genes have diversified during primate evolution despite them conferring a risk of human-specific diseases. To explain the evolutionary process of this diversification, there is a need to focus on mechanisms by which rapid evolutionary changes can occur, such as alternative splicing. Proteins that can bind polyQs are known to act as splicing factors and may provide clues about the rapid evolutionary process. PolyQs are also characterized by the formation of intrinsically disordered (ID) regions, so I hypothesized that polyQs are involved in the transportation of various molecules between the nucleus and cytoplasm to regulate mechanisms characteristic of humans such as neural development. To determine target molecules for empirical research to understand the evolutionary change, I explored protein-protein interactions (PPIs) involving the relevant proteins. This study identified pathways related to polyQ binding as hub proteins scattered across various regulatory systems, including regulation via PQBP1, VCP, or CREBBP. Nine ID hub proteins with both nuclear and cytoplasmic localization were found. Functional annotations suggested that ID proteins containing polyQs are involved in regulating transcription and ubiquitination by flexibly changing PPI formation. These findings explain the relationships among splicing complex, polyQ length variations, and modifications in neural development.

Deep learning-assisted prediction of protein-protein interactions in Arabidopsis thaliana

Currently, the experimentally identified interactome of Arabidopsis (Arabidopsis thaliana) is still far from complete, suggesting that computational prediction methods can complement experimental techniques. Motivated by the prosperity and success of deep learning algorithms and natural language processing techniques, we introduce an integrative deep learning framework, DeepAraPPI, allowing us to predict protein-protein interactions (PPIs) of Arabidopsis utilizing sequence, domain and Gene Ontology (GO) information. Our current DeepAraPPI comprises: (i) a word2vec encoding-based Siamese recurrent convolutional neural network (RCNN) model; (ii) a Domain2vec encoding-based multiple-layer perceptron (MLP) model; and (iii) a GO2vec encoding-based MLP model. Finally, DeepAraPPI combines the prediction results of the three individual predictors through a logistic regression model. Compiling high-quality positive and negative training and test samples by applying strict filtering strategies, DeepAraPPI shows superior performance compared with existing state-of-the-art Arabidopsis PPI prediction methods. DeepAraPPI also provides better cross-species predictive ability in rice (Oryza sativa) than traditional machine learning methods, although the overall performance in cross-species prediction remains to be improved. DeepAraPPI is freely accessible at http://zzdlab.com/deeparappi/. In the meantime, we have also made the source code and data sets of DeepAraPPI available at https://github.com/zjy1125/DeepAraPPI.

SGPPI: structure-aware prediction of protein-protein interactions in rigorous conditions with graph convolutional network

While deep learning (DL)-based models have emerged as powerful approaches to predict protein-protein interactions (PPIs), the reliance on explicit similarity measures (e.g. sequence similarity and network neighborhood) to known interacting proteins makes these methods ineffective in dealing with novel proteins. The advent of AlphaFold2 presents a significant opportunity and also a challenge to predict PPIs in a straightforward way based on monomer structures while controlling bias from protein sequences. In this work, we established Structure and Graph-based Predictions of Protein Interactions (SGPPI), a structure-based DL framework for predicting PPIs, using the graph convolutional network. In particular, SGPPI focused on protein patches on the protein-protein binding interfaces and extracted the structural, geometric and evolutionary features from the residue contact map to predict PPIs. We demonstrated that our model outperforms traditional machine learning methods and state-of-the-art DL-based methods using non-representation-bias benchmark datasets. Moreover, our model trained on human dataset can be reliably transferred to predict yeast PPIs, indicating that SGPPI can capture converging structural features of protein interactions across various species. The implementation of SGPPI is available at https://github.com/emerson106/SGPPI.

Catechol-O-methyl transferase suppresses cell invasion and interplays with MET signaling in estrogen dependent breast cancer

Catechol-O-methyl transferase (COMT) is involved in detoxification of catechol estrogens, playing cancer-protective role in cells producing or utilizing estrogen. Moreover, COMT suppressed migration potential of breast cancer (BC) cells. To delineate COMT role in metastasis of estrogen receptor (ER) dependent BC, we investigated the effect of COMT overexpression on invasion, transcriptome, proteome and interactome of MCF7 cells, a luminal A BC model, stably transduced with lentiviral vector carrying COMT gene (MCF7-COMT). 2D and 3D assays revealed that COMT overexpression associates with decreased cell invasion (p < 0.0001 for Transwell assay, p < 0.05 for spheroid formation). RNA-Seq and LC-DIA-MS/MS proteomics identified genes associated with invasion (FTO, PIR, TACSTD2, ANXA3, KRT80, S100P, PREX1, CLEC3A, LCP1) being downregulated in MCF7-COMT cells, while genes associated with less aggressive phenotype (RBPMS, ROBO2, SELENBP, EPB41L2) were upregulated both at transcript (|log2FC|> 1, adj. p < 0.05) and protein (|log2FC|> 0.58, q < 0.05) levels. Importantly, proteins driving MET signaling were less abundant in COMT overexpressing cells, and pull-down confirmed interaction between COMT and Kunitz-type protease inhibitor 2 (SPINT2), a negative regulator of MET (log2FC = 5.10, q = 1.04-7). In conclusion, COMT may act as tumor suppressor in ER dependent BC not only by detoxification of catechol estrogens but also by suppressing cell invasion and interplay with MET pathway.

A Sequential Targeting Strategy Interrupts AKT-Driven Subclone-Mediated Progression in Glioblastoma

PURPOSE

Therapy resistance and fatal disease progression in glioblastoma are thought to result from the dynamics of intra-tumor heterogeneity. This study aimed at identifying and molecularly targeting tumor cells that can survive, adapt, and subclonally expand under primary therapy.

EXPERIMENTAL DESIGN

To identify candidate markers and to experimentally access dynamics of subclonal progression in glioblastoma, we established a discovery cohort of paired vital cell samples obtained before and after primary therapy. We further used two independent validation cohorts of paired clinical tissues to test our findings. Follow-up preclinical treatment strategies were evaluated in patient-derived xenografts.

RESULTS

We describe, in clinical samples, an archetype of rare ALDH1A1+ tumor cells that enrich and acquire AKT-mediated drug resistance in response to standard-of-care temozolomide (TMZ). Importantly, we observe that drug resistance of ALDH1A1+ cells is not intrinsic, but rather an adaptive mechanism emerging exclusively after TMZ treatment. In patient cells and xenograft models of disease, we recapitulate the enrichment of ALDH1A1+ cells under the influence of TMZ. We demonstrate that their subclonal progression is AKT-driven and can be interfered with by well-timed sequential rather than simultaneous antitumor combination strategy.

CONCLUSIONS

Drug-resistant ALDH1A1+/pAKT+ subclones accumulate in patient tissues upon adaptation to TMZ therapy. These subclones may therefore represent a dynamic target in glioblastoma. Our study proposes the combination of TMZ and AKT inhibitors in a sequential treatment schedule as a rationale for future clinical investigation.

Computational Methods and Deep Learning for Elucidating Protein Interaction Networks

Protein interactions play a critical role in all biological processes, but experimental identification of protein interactions is a time- and resource-intensive process. The advances in next-generation sequencing and multi-omics technologies have greatly benefited large-scale predictions of protein interactions using machine learning methods. A wide range of tools have been developed to predict protein-protein, protein-nucleic acid, and protein-drug interactions. Here, we discuss the applications, methods, and challenges faced when employing the various prediction methods. We also briefly describe ways to overcome the challenges and prospective future developments in the field of protein interaction biology.

Sequence Representations and Their Utility for Predicting Protein-Protein Interactions

Protein-Protein Interactions (PPIs) are a crucial mechanism underpinning the function of the cell. So far, a wide range of machine-learning based methods have been proposed for predicting these relationships. Their success is heavily dependent on the construction of the underlying feature vectors, with most using a set of physico-chemical properties derived from the sequence. Few work directly with the sequence itself. In this paper, we explore the utility of sequence embeddings for predicting protein-protein interactions. We construct a protein pair feature vector by concatenating the embeddings of their constituent sequence. These feature vectors are then used as input to a binary classifier to make predictions. To learn sequence embeddings, we use two established Word2Vec based methods - Seq2Vec and BioVec - and we also introduce a novel feature construction method called SuperVecNW. The embeddings generated through SuperVecNW capture some network information in addition to the contextual information present in the sequences. We test the efficacy of our proposed approach on human and yeast PPI datasets and on three well-known networks: CD9, the Ras-Raf-Mek-Erk-Elk-Srf pathway, and a Wnt-related network. We demonstrate that low dimensional sequence embeddings provide better results than most alternative representations based on physico-chemical properties while offering a far simple approach to feature vector construction.

Protein-protein interaction prediction with deep learning: A comprehensive review

Most proteins perform their biological function by interacting with themselves or other molecules. Thus, one may obtain biological insights into protein functions, disease prevalence, and therapy development by identifying protein-protein interactions (PPI). However, finding the interacting and non-interacting protein pairs through experimental approaches is labour-intensive and time-consuming, owing to the variety of proteins. Hence, protein-protein interaction and protein-ligand binding problems have drawn attention in the fields of bioinformatics and computer-aided drug discovery. Deep learning methods paved the way for scientists to predict the 3-D structure of proteins from genomes, predict the functions and attributes of a protein, and modify and design new proteins to provide desired functions. This review focuses on recent deep learning methods applied to problems including predicting protein functions, protein-protein interaction and their sites, protein-ligand binding, and protein design.

DEMA: a distance-bounded energy-field minimization algorithm to model and layout biomolecular networks with quantitative features

SUMMARY

In biology, graph layout algorithms can reveal comprehensive biological contexts by visually positioning graph nodes in their relevant neighborhoods. A layout software algorithm/engine commonly takes a set of nodes and edges and produces layout coordinates of nodes according to edge constraints. However, current layout engines normally do not consider node, edge or node-set properties during layout and only curate these properties after the layout is created. Here, we propose a new layout algorithm, distance-bounded energy-field minimization algorithm (DEMA), to natively consider various biological factors, i.e., the strength of gene-to-gene association, the gene's relative contribution weight and the functional groups of genes, to enhance the interpretation of complex network graphs. In DEMA, we introduce a parameterized energy model where nodes are repelled by the network topology and attracted by a few biological factors, i.e., interaction coefficient, effect coefficient and fold change of gene expression. We generalize these factors as gene weights, protein-protein interaction weights, gene-to-gene correlations and the gene set annotations-four parameterized functional properties used in DEMA. Moreover, DEMA considers further attraction/repulsion/grouping coefficient to enable different preferences in generating network views. Applying DEMA, we performed two case studies using genetic data in autism spectrum disorder and Alzheimer's disease, respectively, for gene candidate discovery. Furthermore, we implement our algorithm as a plugin to Cytoscape, an open-source software platform for visualizing networks; hence, it is convenient. Our software and demo can be freely accessed at http://discovery.informatics.uab.edu/dema.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A review of bioinformatics tools and web servers in different microarray platforms used in cancer research

Over the past decade, conventional lab work strategies have gradually shifted from being limited to a laboratory setting towards a bioinformatics era to help manage and process the vast amounts of data generated by omics technologies. The present work outlines the latest contributions of bioinformatics in analyzing microarray data and their application to cancer. We dissect different microarray platforms and their use in gene expression in cancer models. We highlight how computational advances empowered the microarray technology in gene expression analysis. The study on protein-protein interaction databases classified into primary, derived, meta-database, and prediction databases describes the strategies to curate and predict novel interaction networks in silico. In addition, we summarize the areas of bioinformatics where neural graph networks are currently being used, such as protein functions, protein interaction prediction, and in silico drug discovery and development. We also discuss the role of deep learning as a potential tool in the prognosis, diagnosis, and treatment of cancer. Integrating these resources efficiently, practically, and ethically is likely to be the most challenging task for the healthcare industry over the next decade; however, we believe that it is achievable in the long term.

In vitro ve in silico analizi ile metforminin meme tümörü hücrelerinde protein profili üzerindeki etkinliği

Aim: This study aimed to uncover the varieties in protein profiles of Met in breast tumor (BT) cells by assessment of in vitro and in silico analysis. Materials and Methods: Here, the cells obtained from mastectomy patients were cultured, the effective Met-dose was determined as 25 mM through cell viability and BrdU tests. Protein identification in the breast tumor cells was implemented by employing LC-MS/MS technology. Results: The expression of SSR3, THAP3, FTH1, NEFM, ANP32A, ANP32B, KRT7 proteins was significantly decreased whereas the GARS protein increased in the 25 mM Met group compared to the Non-Met (0 mM) control group. In silico analysis, we analyzed the probable interactions of all these proteins with each other and other proteins, to evaluate the analysis of the larger protein network, and which metabolic pathway proteins are involved in. Conclusion: The stated proteomics analysis in our study proposes a better understanding of the prognosis of breast cancer and future studies to investigate the effect of metformin in this field on proteomic pathways in other sorts of cancer.

In silico Methods for Identification of Potential Therapeutic Targets

At the initial stage of drug discovery, identifying novel targets with maximal efficacy and minimal side effects can improve the success rate and portfolio value of drug discovery projects while simultaneously reducing cycle time and cost. However, harnessing the full potential of big data to narrow the range of plausible targets through existing computational methods remains a key issue in this field. This paper reviews two categories of in silico methods—comparative genomics and network-based methods—for finding potential therapeutic targets among cellular functions based on understanding their related biological processes. In addition to describing the principles, databases, software, and applications, we discuss some recent studies and prospects of the methods. While comparative genomics is mostly applied to infectious diseases, network-based methods can be applied to infectious and non-infectious diseases. Nonetheless, the methods often complement each other in their advantages and disadvantages. The information reported here guides toward improving the application of big data-driven computational methods for therapeutic target discovery.

Graphical abstract

Analysis of Huntington's Disease Modifiers Using the Hyperbolic Mapping of the Protein Interaction Network

Huntington's disease (HD) is caused by the production of a mutant huntingtin (HTT) with an abnormally long poly-glutamine (polyQ) tract, forming aggregates and inclusions in neurons. Previous work by us and others has shown that an increase or decrease in polyQ-triggered aggregates can be passive simply due to the interaction of proteins with the aggregates. To search for proteins with active (functional) effects, which might be more effective in finding therapies and mechanisms of HD, we selected among the proteins that interact with HTT a total of 49 pairs of proteins that, while being paralogous to each other (and thus expected to have similar passive interaction with HTT), are located in different regions of the protein interaction network (suggesting participation in different pathways or complexes). Three of these 49 pairs contained members with opposite effects on HD, according to the literature. The negative members of the three pairs, MID1, IKBKG, and IKBKB, interact with PPP2CA and TUBB, which are known negative factors in HD, as well as with HSP90AA1 and RPS3. The positive members of the three pairs interact with HSPA9. Our results provide potential HD modifiers of functional relevance and reveal the dynamic aspect of paralog evolution within the interaction network.

Proteome-wide prediction and analysis of the Cryptosporidium parvum protein-protein interaction network through integrative methods

•

By combining a sequence embedding technique (i.e., Doc2Vec) and a di-peptide composition representation to convert protein sequences into feature vectors, we proposed an RF classifier trained on the Plasmodium falciparum dataset for predicting Cryptosporidium parvum PPIs.

•

A high-confidence Cryptosporidium parvum PPI network was identified by conjoining interolog mapping, domain-domain interaction-based inference, and the RF classifier.

•

Some detected hub proteins and functional modules provided clues for an in-depth biological understanding of Cryptosporidium parvum.

As one of the most studied Apicomplexan parasite Cryptosporidium, Cryptosporidium parvum (C. parvum) causes worldwide serious diarrhea disease cryptosporidiosis, which can be deadly to immunodeficiency individuals, newly born children, and animals. Proteome-wide identification of protein–protein interactions (PPIs) has proven valuable in the systematic understanding of the genome-phenome relationship. However, the PPIs of C. parvum are largely unknown because of the limited experimental studies carried out. Therefore, we took full advantage of three bioinformatics methods, i.e., interolog mapping (IM), domain-domain interaction (DDI)-based inference, and machine learning (ML) method, to jointly predict PPIs of C. parvum. Due to the lack of experimental PPIs of C. parvum, we used the PPI data of Plasmodium falciparum (P. falciparum), which owned the largest number of PPIs in Apicomplexa, to train an ML model to infer C. parvum PPIs. We utilized consistent results of these three methods as the predicted high-confidence PPI network, which contains 4,578 PPIs covering 554 proteins. To further explore the biological significance of the constructed PPI network, we also conducted essential network and protein functional analysis, mainly focusing on hub proteins and functional modules. We anticipate the constructed PPI network can become an important data resource to accelerate the functional genomics studies of C. parvum as well as offer new hints to the target discovery in developing drugs/vaccines.

S1P defects cause a new entity of cataract, alopecia, oral mucosal disorder, and psoriasis-like syndrome

In this report, we discovered a new entity named cataract, alopecia, oral mucosal disorder, and psoriasis‐like (CAOP) syndrome in two unrelated and ethnically diverse patients. Furthermore, patient 1 failed to respond to regular treatment. We found that CAOP syndrome was caused by an autosomal recessive defect in the mitochondrial membrane‐bound transcription factor peptidase/site‐1 protease (MBTPS1, S1P). Mitochondrial abnormalities were observed in patient 1 with CAOP syndrome. Furthermore, we found that S1P is a novel mitochondrial protein that forms a trimeric complex with ETFA/ETFB. S1P enhances ETFA/ETFB flavination and maintains its stability. Patient S1P variants destabilize ETFA/ETFB, impair mitochondrial respiration, decrease fatty acid β‐oxidation activity, and shift mitochondrial oxidative phosphorylation (OXPHOS) to glycolysis. Mitochondrial dysfunction and inflammatory lesions in patient 1 were significantly ameliorated by riboflavin supplementation, which restored the stability of ETFA/ETFB. Our study discovered that mutations in MBTPS1 resulted in a new entity of CAOP syndrome and elucidated the mechanism of the mutations in the new disease.

Machine learning methods for prediction of cancer driver genes: a survey paper

Identifying the genes and mutations that drive the emergence of tumors is a critical step to improving our understanding of cancer and identifying new directions for disease diagnosis and treatment. Despite the large volume of genomics data, the precise detection of driver mutations and their carrying genes, known as cancer driver genes, from the millions of possible somatic mutations remains a challenge. Computational methods play an increasingly important role in discovering genomic patterns associated with cancer drivers and developing predictive models to identify these elements. Machine learning (ML), including deep learning, has been the engine behind many of these efforts and provides excellent opportunities for tackling remaining gaps in the field. Thus, this survey aims to perform a comprehensive analysis of ML-based computational approaches to identify cancer driver mutations and genes, providing an integrated, panoramic view of the broad data and algorithmic landscape within this scientific problem. We discuss how the interactions among data types and ML algorithms have been explored in previous solutions and outline current analytical limitations that deserve further attention from the scientific community. We hope that by helping readers become more familiar with significant developments in the field brought by ML, we may inspire new researchers to address open problems and advance our knowledge towards cancer driver discovery.

Integrative system biology and mathematical modeling of genetic networks identifies shared biomarkers for obesity and diabetes

Obesity and type 2 and diabetes mellitus (T2D) are two dual epidemics whose shared genetic pathological mechanisms are still far from being fully understood. Therefore, this study is aimed at discovering key genes, molecular mechanisms, and new drug targets for obesity and T2D by analyzing the genome wide gene expression data with different computational biology approaches. In this study, the RNA-sequencing data of isolated primary human adipocytes from individuals who are lean, obese, and T2D was analyzed by an integrated framework consisting of gene expression, protein interaction network (PIN), tissue specificity, and druggability approaches. Our findings show a total of 1932 unique differentially expressed genes (DEGs) across the diabetes versus obese group comparison (p≤0.05). The PIN analysis of these 1932 DEGs identified 190 high centrality network (HCN) genes, which were annotated against 3367 GO terms and functional pathways, like response to insulin signaling, phosphorylation, lipid metabolism, glucose metabolism, etc. (p≤0.05). By applying additional PIN and topological parameters to 190 HCN genes, we further mapped 25 high confidence genes, functionally connected with diabetes and obesity traits. Interestingly, ERBB2, FN1, FYN, HSPA1A, HBA1, and ITGB1 genes were found to be tractable by small chemicals, antibodies, and/or enzyme molecules. In conclusion, our study highlights the potential of computational biology methods in correlating expression data to topological parameters, functional relationships, and druggability characteristics of the candidate genes involved in complex metabolic disorders with a common etiological basis.

From complete cross-docking to partners identification and binding sites predictions

Proteins ensure their biological functions by interacting with each other. Hence, characterising protein interactions is fundamental for our understanding of the cellular machinery, and for improving medicine and bioengineering. Over the past years, a large body of experimental data has been accumulated on who interacts with whom and in what manner. However, these data are highly heterogeneous and sometimes contradictory, noisy, and biased. Ab initio methods provide a means to a "blind" protein-protein interaction network reconstruction. Here, we report on a molecular cross-docking-based approach for the identification of protein partners. The docking algorithm uses a coarse-grained representation of the protein structures and treats them as rigid bodies. We applied the approach to a few hundred of proteins, in the unbound conformations, and we systematically investigated the influence of several key ingredients, such as the size and quality of the interfaces, and the scoring function. We achieved some significant improvement compared to previous works, and a very high discriminative power on some specific functional classes. We provide a readout of the contributions of shape and physico-chemical complementarity, interface matching, and specificity, in the predictions. In addition, we assessed the ability of the approach to account for protein surface multiple usages, and we compared it with a sequence-based deep learning method. This work may contribute to guiding the exploitation of the large amounts of protein structural models now available toward the discovery of unexpected partners and their complex structure characterisation.

A New Family of Similarity Measures for Scoring Confidence of Protein Interactions Using Gene Ontology

The large-scale protein-protein interaction (PPI) data has the potential to play a significant role in the endeavor of understanding cellular processes. However, the presence of a considerable fraction of false positives is a bottleneck in realizing this potential. There have been continuous efforts to utilize complementary resources for scoring confidence of PPIs in a manner that false positive interactions get a low confidence score. Gene Ontology (GO), a taxonomy of biological terms to represent the properties of gene products and their relations, has been widely used for this purpose. We utilize GO to introduce a new set of specificity measures: Relative Depth Specificity (RDS), Relative Node-based Specificity (RNS), and Relative Edge-based Specificity (RES), leading to a new family of similarity measures. We use these similarity measures to obtain a confidence score for each PPI. We evaluate the new measures using four different benchmarks. We show that all the three measures are quite effective. Notably, RNS and RES more effectively distinguish true PPIs from false positives than the existing alternatives. RES also shows a robust set-discriminating power and can be useful for protein functional clustering as well.

Testing biological network motif significance with exponential random graph models

Analysis of the structure of biological networks often uses statistical tests to establish the over-representation of motifs, which are thought to be important building blocks of such networks, related to their biological functions. However, there is disagreement as to the statistical significance of these motifs, and there are potential problems with standard methods for estimating this significance. Exponential random graph models (ERGMs) are a class of statistical model that can overcome some of the shortcomings of commonly used methods for testing the statistical significance of motifs. ERGMs were first introduced into the bioinformatics literature over 10 years ago but have had limited application to biological networks, possibly due to the practical difficulty of estimating model parameters. Advances in estimation algorithms now afford analysis of much larger networks in practical time. We illustrate the application of ERGM to both an undirected protein-protein interaction (PPI) network and directed gene regulatory networks. ERGM models indicate over-representation of triangles in the PPI network, and confirm results from previous research as to over-representation of transitive triangles (feed-forward loop) in an E. coli and a yeast regulatory network. We also confirm, using ERGMs, previous research showing that under-representation of the cyclic triangle (feedback loop) can be explained as a consequence of other topological features.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s41109-021-00434-y.

Tumor relevant protein functional interactions identified using bipartite graph analyses

An increased surge of -omics data for the diseases such as cancer allows for deriving insights into the affiliated protein interactions. We used bipartite network principles to build protein functional associations of the differentially regulated genes in 18 cancer types. This approach allowed us to combine expression data to functional associations in many cancers simultaneously. Further, graph centrality measures suggested the importance of upregulated genes such as BIRC5, UBE2C, BUB1B, KIF20A and PTH1R in cancer. Pathway analysis of the high centrality network nodes suggested the importance of the upregulation of cell cycle and replication associated proteins in cancer. Some of the downregulated high centrality proteins include actins, myosins and ATPase subunits. Among the transcription factors, mini-chromosome maintenance proteins (MCMs) and E2F family proteins appeared prominently in regulating many differentially regulated genes. The projected unipartite networks of the up and downregulated genes were comprised of 37,411 and 41,756 interactions, respectively. The conclusions obtained by collating these interactions revealed pan-cancer as well as subtype specific protein complexes and clusters. Therefore, we demonstrate that incorporating expression data from multiple cancers into bipartite graphs validates existing cancer associated mechanisms as well as directs to novel interactions and pathways.

Performance Assessment of the Network Reconstruction Approaches on Various Interactomes

Beyond the list of molecules, there is a necessity to collectively consider multiple sets of omic data and to reconstruct the connections between the molecules. Especially, pathway reconstruction is crucial to understanding disease biology because abnormal cellular signaling may be pathological. The main challenge is how to integrate the data together in an accurate way. In this study, we aim to comparatively analyze the performance of a set of network reconstruction algorithms on multiple reference interactomes. We first explored several human protein interactomes, including PathwayCommons, OmniPath, HIPPIE, iRefWeb, STRING, and ConsensusPathDB. The comparison is based on the coverage of each interactome in terms of cancer driver proteins, structural information of protein interactions, and the bias toward well-studied proteins. We next used these interactomes to evaluate the performance of network reconstruction algorithms including all-pair shortest path, heat diffusion with flux, personalized PageRank with flux, and prize-collecting Steiner forest (PCSF) approaches. Each approach has its own merits and weaknesses. Among them, PCSF had the most balanced performance in terms of precision and recall scores when 28 pathways from NetPath were reconstructed using the listed algorithms. Additionally, the reference interactome affects the performance of the network reconstruction approaches. The coverage and disease- or tissue-specificity of each interactome may vary, which may result in differences in the reconstructed networks.

In silico comparative analysis of LRRK2 interactomes from brain, kidney and lung

Mutations in LRRK2 are the most frequent cause of familial Parkinson's disease (PD), with common LRRK2 non-coding variants also acting as risk factors for idiopathic PD. Currently, therapeutic agents targeting LRRK2 are undergoing advanced clinical trials in humans, however, it is important to understand the wider implications of LRRK2 targeted treatments given that LRRK2 is expressed in diverse tissues including the brain, kidney and lungs. This presents challenges to treatment in terms of effects on peripheral organ functioning, thus, protein interactors of LRRK2 could be targeted in lieu to optimize therapeutic effects. Herein an in-silico analysis of LRRK2 direct interactors in brain tissue from various brain regionswas conducted along with a comparative analysis of the LRRK2 interactome in the brain, kidney, and lung tissues. This was carried out based on curated protein-protein interaction (PPI) data from protein interaction databases such as HIPPIE, human gene/protein expression databases and Gene ontology (GO) enrichment analysis using Bingo. Seven targets (MAP2K6, MATK, MAPT, PAK6, SH3GL2, CDC42EP3 and CHGB) were found to be viable objectives for LRRK2 based investigations for PD that would have minimal impact on optimal functioning within peripheral organs. Specifically, MAPT, CHGB, PAK6, and SH3GL2 interacted with LRRK2 in the brain and kidney but not in lung tissue whilst LRRK2-MAP2K6 interacted only in the cerebellum and MATK-LRRK2 interaction was absent in kidney tissues. CDC42EP3 expression levels were low in brain tissues compared to kidney/lung. The results of this computational analysis suggest new avenues for experimental investigations towards LRRK2-targeted therapeutics.

Matrix-screening reveals a vast potential for direct protein-protein interactions among RNA binding proteins

RNA-binding proteins (RBPs) are crucial factors of post-transcriptional gene regulation and their modes of action are intensely investigated. At the center of attention are RNA motifs that guide where RBPs bind. However, sequence motifs are often poor predictors of RBP-RNA interactions in vivo. It is hence believed that many RBPs recognize RNAs as complexes, to increase specificity and regulatory possibilities. To probe the potential for complex formation among RBPs, we assembled a library of 978 mammalian RBPs and used rec-Y2H matrix screening to detect direct interactions between RBPs, sampling > 600 K interactions. We discovered 1994 new interactions and demonstrate that interacting RBPs bind RNAs adjacently in vivo. We further find that the mRNA binding region and motif preferences of RBPs deviate, depending on their adjacently binding interaction partners. Finally, we reveal novel RBP interaction networks among major RNA processing steps and show that splicing impairing RBP mutations observed in cancer rewire spliceosomal interaction networks. The dataset we provide will be a valuable resource for understanding the combinatorial interactions of RBPs with RNAs and the resulting regulatory outcomes.

MIPPIE: the mouse integrated protein-protein interaction reference

Cells operate and react to environmental signals thanks to a complex network of protein–protein interactions (PPIs), the malfunction of which can severely disrupt cellular homeostasis. As a result, mapping and analyzing protein networks are key to advancing our understanding of biological processes and diseases. An invaluable part of these endeavors has been the house mouse (Mus musculus), the mammalian model organism par excellence, which has provided insights into human biology and disorders. The importance of investigating PPI networks in the context of mouse prompted us to develop the Mouse Integrated Protein–Protein Interaction rEference (MIPPIE). MIPPIE inherits a robust infrastructure from HIPPIE, its sister database of human PPIs, allowing for the assembly of reliable networks supported by different evidence sources and high-quality experimental techniques. MIPPIE networks can be further refined with tissue, directionality and effect information through a user-friendly web interface. Moreover, all MIPPIE data and meta-data can be accessed via a REST web service or downloaded as text files, thus facilitating the integration of mouse PPIs into follow-up bioinformatics pipelines.

Bioinformatic analysis of the human brain extracellular matrix proteome in neurodegenerative disorders

Alzheimer's, Parkinson's, and Huntington's diseases are characterized by selective degeneration of specific brain areas. Although increasing number of studies report alteration of the extracellular matrix on these diseases, an exhaustive characterization at the brain's matrix level might contribute to the development of more efficient cell restoration therapies. In that regard, proteomics-based studies are a powerful approach to uncover matrix changes. However, to date, the majority of proteomics studies report no or only a few brain matrix proteins with altered expression. This study aims to reveal the changes in the brain extracellular matrix by integrating several proteomics-based studies performed with postmortem tissue. In total, 67 matrix proteins with altered expression were collected. By applying a bioinformatic approach, we were able to reveal the dysregulated biological processes. Among them are processes related to the organization of the extracellular matrix, glycosaminoglycans and proteoglycans' metabolism, blood coagulation, and response to injury and oxidative stress. In addition, a protein was found altered in all three diseases-collagen type I alpha 2-and its binding partners further identified. A ClueGO network was created, depicting the GO groups associated with these binding partners, uncovering the processes that may consequently be affected. These include cellular adhesion, cell signaling through membrane receptors, inflammatory processes, and apoptotic cell death in response to oxidative stress. Overall, we were able to associate the contribution of the modification of extracellular matrix components to essential biological processes, highlighting the investment needed on proteomics studies with specific focus on the extracellular matrix in neurodegeneration.

Predicting Protein-Protein Interactions via Gated Graph Attention Signed Network

Protein-protein interactions (PPIs) play a key role in signal transduction and pharmacogenomics, and hence, accurate PPI prediction is crucial. Graph structures have received increasing attention owing to their outstanding performance in machine learning. In practice, PPIs can be expressed as a signed network (i.e., graph structure), wherein the nodes in the network represent proteins, and edges represent the interactions (positive or negative effects) of protein nodes. PPI predictions can be realized by predicting the links of the signed network; therefore, the use of gated graph attention for signed networks (SN-GGAT) is proposed herein. First, the concept of graph attention network (GAT) is applied to signed networks, in which "attention" represents the weight of neighbor nodes, and GAT updates the node features through the weighted aggregation of neighbor nodes. Then, the gating mechanism is defined and combined with the balance theory to obtain the high-order relations of protein nodes to improve the attention effect, making the attention mechanism follow the principle of "low-order high attention, high-order low attention, different signs opposite". PPIs are subsequently predicted on the Saccharomyces cerevisiae core dataset and the Human dataset. The test results demonstrate that the proposed method exhibits strong competitiveness.

Predicting tissue-specific gene expression from whole blood transcriptome

Complex diseases are mediated via transcriptional dysregulation in multiple tissues. Thus, knowing an individual's tissue-specific gene expression can provide critical information about her health. Unfortunately, for most tissues, the transcriptome cannot be obtained without invasive procedures. Could we, however, infer an individual's tissue-specific expression from her whole blood transcriptome? Here, we rigorously address this question. We find that an individual's whole blood transcriptome can significantly predict tissue-specific expression levels for ~60% of the genes on average across 32 tissues, with up to 81% of the genes in skeletal muscle. The tissue-specific expression inferred from the blood transcriptome is almost as good as the actual measured tissue expression in predicting disease state for six different complex disorders, including hypertension and type 2 diabetes, substantially surpassing the blood transcriptome. The code for tissue-specific gene expression prediction, TEEBoT, is provided, enabling others to study its potential translational value in other indications.

Biological network analysis with deep learning

Recent advancements in experimental high-throughput technologies have expanded the availability and quantity of molecular data in biology. Given the importance of interactions in biological processes, such as the interactions between proteins or the bonds within a chemical compound, this data is often represented in the form of a biological network. The rise of this data has created a need for new computational tools to analyze networks. One major trend in the field is to use deep learning for this goal and, more specifically, to use methods that work with networks, the so-called graph neural networks (GNNs). In this article, we describe biological networks and review the principles and underlying algorithms of GNNs. We then discuss domains in bioinformatics in which graph neural networks are frequently being applied at the moment, such as protein function prediction, protein-protein interaction prediction and in silico drug discovery and development. Finally, we highlight application areas such as gene regulatory networks and disease diagnosis where deep learning is emerging as a new tool to answer classic questions like gene interaction prediction and automatic disease prediction from data.

Biological network analysis with deep learning

Recent advancements in experimental high-throughput technologies have expanded the availability and quantity of molecular data in biology. Given the importance of interactions in biological processes, such as the interactions between proteins or the bonds within a chemical compound, this data is often represented in the form of a biological network. The rise of this data has created a need for new computational tools to analyze networks. One major trend in the field is to use deep learning for this goal and, more specifically, to use methods that work with networks, the so-called graph neural networks (GNNs). In this article, we describe biological networks and review the principles and underlying algorithms of GNNs. We then discuss domains in bioinformatics in which graph neural networks are frequently being applied at the moment, such as protein function prediction, protein-protein interaction prediction and in silico drug discovery and development. Finally, we highlight application areas such as gene regulatory networks and disease diagnosis where deep learning is emerging as a new tool to answer classic questions like gene interaction prediction and automatic disease prediction from data.

A genome-wide association study identifies a gene network associated with paranoid schizophrenia and antipsychotics-induced tardive dyskinesia

In the present study we conducted a genome-wide association study (GWAS) in a cohort of 505 patients with paranoid schizophrenia (SCZ), of which 95 had tardive dyskinesia (TD), and 503 healthy controls. Using data generated by the PsychENCODE Consortium (PEC) and other bioinformatic databases, we revealed a gene network, implicated in neurodevelopment and brain function, associated with both these disorders. Almost all these genes are in gene or isoform co-expression PEC network modules important for the functioning of the brain; the activity of these networks is also altered in SCZ, bipolar disorder and autism spectrum disorders. The associated PEC network modules are enriched for gene ontology terms relevant to the brain development and function (CNS development, neuron development, axon ensheathment, synapse, synaptic vesicle cycle, and signaling receptor activity) and to the immune system (inflammatory response). Results of the present study suggest that orofacial and limbtruncal types of TD seem to share the molecular network with SCZ. Paranoid SCZ and abnormal involuntary movements that indicate the orofacial type of TD are associated with the same genomic loci on chromosomes 3p22.2, 8q21.13, and 13q14.2. The limbtruncal type of TD is associated with a locus on chromosome 3p13 where the best functional candidate is FOXP1, a high-confidence SCZ gene. The results of this study shed light on common pathogenic mechanisms for SCZ and TD, and indicate that the pathogenesis of the orofacial and limbtruncal types of TD might be driven by interacting genes implicated in neurodevelopment.

Predicting disease‐associated genes: Computational methods, databases, and evaluations

Complex diseases are associated with a set of genes (called disease genes), the identification of which can help scientists uncover the mechanisms of diseases and develop new drugs and treatment strategies. Due to the huge cost and time of experimental identification techniques, many computational algorithms have been proposed to predict disease genes. Although several review publications in recent years have discussed many computational methods, some of them focus on cancer driver genes while others focus on biomolecular networks, which only cover a specific aspect of existing methods. In this review, we summarize existing methods and classify them into three categories based on their rationales. Then, the algorithms, biological data, and evaluation methods used in the computational prediction are discussed. Finally, we highlight the limitations of existing methods and point out some future directions for improving these algorithms. This review could help investigators understand the principles of existing methods, and thus develop new methods to advance the computational prediction of disease genes.

This article is categorized under: