Duplicability of self-interacting human genes

NLPEI: A Novel Self-Interacting Protein Prediction Model Based on Natural Language Processing and Evolutionary Information

The study of protein self-interactions (SIPs) can not only reveal the function of proteins at the molecular level, but is also crucial to understand activities such as growth, development, differentiation, and apoptosis, providing an important theoretical basis for exploring the mechanism of major diseases. With the rapid advances in biotechnology, a large number of SIPs have been discovered. However, due to the long period and high cost inherent to biological experiments, the gap between the identification of SIPs and the accumulation of data is growing. Therefore, fast and accurate computational methods are needed to effectively predict SIPs. In this study, we designed a new method, NLPEI, for predicting SIPs based on natural language understanding theory and evolutionary information. Specifically, we first understand the protein sequence as natural language and use natural language processing algorithms to extract its features. Then, we use the Position-Specific Scoring Matrix (PSSM) to represent the evolutionary information of the protein and extract its features through the Stacked Auto-Encoder (SAE) algorithm of deep learning. Finally, we fuse the natural language features of proteins with evolutionary features and make accurate predictions by Extreme Learning Machine (ELM) classifier. In the SIPs gold standard data sets of human and yeast, NLPEI achieved 94.19% and 91.29% prediction accuracy. Compared with different classifier models, different feature models, and other existing methods, NLPEI obtained the best results. These experimental results indicated that NLPEI is an effective tool for predicting SIPs and can provide reliable candidates for biological experiments.

Multivalent Proteins Rapidly and Reversibly Phase-Separate upon Osmotic Cell Volume Change

Processing bodies (PBs) and stress granules (SGs) are prominent examples of subcellular, membraneless compartments that are observed under physiological and stress conditions, respectively. We observe that the trimeric PB protein DCP1A rapidly (within ∼10 s) phase-separates in mammalian cells during hyperosmotic stress and dissolves upon isosmotic rescue (over ∼100 s) with minimal effect on cell viability even after multiple cycles of osmotic perturbation. Strikingly, this rapid intracellular hyperosmotic phase separation (HOPS) correlates with the degree of cell volume compression, distinct from SG assembly, and is exhibited broadly by homo-multimeric (valency ≥ 2) proteins across several cell types. Notably, HOPS sequesters pre-mRNA cleavage factor components from actively transcribing genomic loci, providing a mechanism for hyperosmolarity-induced global impairment of transcription termination. Our data suggest that the multimeric proteome rapidly responds to changes in hydration and molecular crowding, revealing an unexpected mode of globally programmed phase separation and sequestration.

Tetraspanin 7 and its closest paralog tetraspanin 6: membrane organizers with key functions in brain development, viral infection, innate immunity, diabetes and cancer

Tetraspanin (TSPAN) protein family forms a family of transmembrane proteins that act as organizers/scaffold for other proteins. TSPANs are primarily present on plasma membranes although they are also found in other biological membranes. They are organized in tetraspanin-enriched microdomains (TEMs), which allow spatiotemporal tuning of protein functions through the control of their membrane localization. TSPAN6 and TSPAN7 are close paralogs expressed in different tissues, TSPAN7 being highly expressed in the brain. Their functions only started to be unveiled in the late 2000's and are still poorly understood. Here, we introduce how TSPAN7 was first highlighted has a protein mutated in some forms of X-linked mental retardation, which was later proposed to be caused by defects in neuronal morphogenesis and synaptic transmission. We then discuss the impacts TSPAN7 has on cell morphology of dendritic cells and osteoclasts, through rearrangement of actin cytoskeleton and how TSPAN7 was shown to be a target of autoantibody in patients suffering from type 1 diabetes. Finally, we are addressing the double edge sword that is TSPAN7 in cancer. In the second part of this review, we address the known roles of TSPAN6 and how this protein was shown to participate in synaptic transmission and in amyloid precursor protein secretion, which may contribute to Alzheimer's disease pathology. We conclude this review by discussing the anti-inflammatory effect of TSPAN6.

The role of structural pleiotropy and regulatory evolution in the retention of heteromers of paralogs

Gene duplication is a driver of the evolution of new functions. The duplication of genes encoding homomeric proteins leads to the formation of homomers and heteromers of paralogs, creating new complexes after a single duplication event. The loss of these heteromers may be required for the two paralogs to evolve independent functions. Using yeast as a model, we find that heteromerization is frequent among duplicated homomers and correlates with functional similarity between paralogs. Using in silico evolution, we show that for homomers and heteromers sharing binding interfaces, mutations in one paralog can have structural pleiotropic effects on both interactions, resulting in highly correlated responses of the complexes to selection. Therefore, heteromerization could be preserved indirectly due to selection for the maintenance of homomers, thus slowing down functional divergence between paralogs. We suggest that paralogs can overcome the obstacle of structural pleiotropy by regulatory evolution at the transcriptional and post-translational levels.

An Improved Deep Forest Model for Predicting Self-Interacting Proteins From Protein Sequence Using Wavelet Transformation

Self-interacting proteins (SIPs), whose more than two identities can interact with each other, play significant roles in the understanding of cellular process and cell functions. Although a number of experimental methods have been designed to detect the SIPs, they remain to be extremely time-consuming, expensive, and challenging even nowadays. Therefore, there is an urgent need to develop the computational methods for predicting SIPs. In this study, we propose a deep forest based predictor for accurate prediction of SIPs using protein sequence information. More specifically, a novel feature representation method, which integrate position-specific scoring matrix (PSSM) with wavelet transform, is introduced. To evaluate the performance of the proposed method, cross-validation tests are performed on two widely used benchmark datasets. The experimental results show that the proposed model achieved high accuracies of 95.43 and 93.65% on human and yeast datasets, respectively. The AUC value for evaluating the performance of the proposed method was also reported. The AUC value for yeast and human datasets are 0.9203 and 0.9586, respectively. To further show the advantage of the proposed method, it is compared with several existing methods. The results demonstrate that the proposed model is better than other SIPs prediction methods. This work can offer an effective architecture to biologists in detecting new SIPs.

Genome-wide identification and characterization of CKIN/SnRK gene family in Chlamydomonas reinhardtii

The SnRK (Snf1-Related protein Kinase) gene family plays an important role in energy sensing and stress-adaptive responses in plant systems. In this study, Chlamydomonas CKIN family (SnRK in Arabidopsis) was defined after a genome-wide analysis of all sequenced Chlorophytes. Twenty-two sequences were defined as plant SnRK orthologs in Chlamydomonas and classified into two subfamilies: CKIN1 and CKIN2. While CKIN1 subfamily is reduced to one conserved member and a close protein (CKIN1L), a large CKIN2 subfamily clusters both plant-like and algae specific CKIN2s. The responsiveness of these genes to abiotic stress situations was tested by RT-qPCR. Results showed that almost all elements were sensitive to osmotic stress while showing different degrees of sensibility to other abiotic stresses, as occurs in land plants, revealing their specialization and the family pleiotropy for some elements. The regulatory pathway of this family may differ from land plants since these sequences shows unique regulatory features and some of them are sensitive to ABA, despite conserved ABA receptors (PYR/PYL/RCAR) and regulatory domains are not present in this species. Core Chlorophytes and land plant showed divergent stress signalling, but SnRKs/CKINs share the same role in cell survival and stress response and adaption including the accumulation of specific biomolecules. This fact places the CKIN family as well-suited target for bioengineering-based studies in microalgae (accumulation of sugars, lipids, secondary metabolites), while promising new findings in stress biology and specially in the evolution of ABA-signalling mechanisms.

Recent advances in understanding the roles of whole genome duplications in evolution

Ancient whole-genome duplications (WGDs)- paleopolyploidy events-are key to solving Darwin's 'abominable mystery' of how flowering plants evolved and radiated into a rich variety of species. The vertebrates also emerged from their invertebrate ancestors via two WGDs, and genomes of diverse gymnosperm trees, unicellular eukaryotes, invertebrates, fishes, amphibians and even a rodent carry evidence of lineage-specific WGDs. Modern polyploidy is common in eukaryotes, and it can be induced, enabling mechanisms and short-term cost-benefit assessments of polyploidy to be studied experimentally. However, the ancient WGDs can be reconstructed only by comparative genomics: these studies are difficult because the DNA duplicates have been through tens or hundreds of millions of years of gene losses, mutations, and chromosomal rearrangements that culminate in resolution of the polyploid genomes back into diploid ones (rediploidisation). Intriguing asymmetries in patterns of post-WGD gene loss and retention between duplicated sets of chromosomes have been discovered recently, and elaborations of signal transduction systems are lasting legacies from several WGDs. The data imply that simpler signalling pathways in the pre-WGD ancestors were converted via WGDs into multi-stranded parallelised networks. Genetic and biochemical studies in plants, yeasts and vertebrates suggest a paradigm in which different combinations of sister paralogues in the post-WGD regulatory networks are co-regulated under different conditions. In principle, such networks can respond to a wide array of environmental, sensory and hormonal stimuli and integrate them to generate phenotypic variety in cell types and behaviours. Patterns are also being discerned in how the post-WGD signalling networks are reconfigured in human cancers and neurological conditions. It is fascinating to unpick how ancient genomic events impact on complexity, variety and disease in modern life.

SPAR: a random forest-based predictor for self-interacting proteins with fine-grained domain information

Protein self-interaction, i.e. the interaction between two or more identical proteins expressed by one gene, plays an important role in the regulation of cellular functions. Considering the limitations of experimental self-interaction identification, it is necessary to design specific bioinformatics tools for self-interacting protein (SIP) prediction from protein sequence information. In this study, we proposed an improved computational approach for SIP prediction, termed SPAR (Self-interacting Protein Analysis serveR). Firstly, we developed an improved encoding scheme named critical residues substitution (CRS), in which the fine-grained domain-domain interaction information was taken into account. Then, by employing the Random Forest algorithm, the performance of CRS was evaluated and compared with several other encoding schemes commonly used for sequence-based protein-protein interaction prediction. Through the tenfold cross-validation tests on a balanced training dataset, CRS performed the best, with the average accuracy up to 72.01 %. We further integrated CRS with other encoding schemes and identified the most important features using the mRMR (the minimum redundancy maximum relevance) feature selection method. Our SPAR model with selected features achieved an average accuracy of 92.09 % on the human-independent test set (the ratio of positives to negatives was about 1:11). Besides, we also evaluated the performance of SPAR on an independent yeast test set (the ratio of positives to negatives was about 1:8) and obtained an average accuracy of 76.96 %. The results demonstrate that SPAR is capable of achieving a reasonable performance in cross-species application. The SPAR server is freely available for academic use at http://systbio.cau.edu.cn/zzdlab/spar/ .

Patterns of Gene Conversion in Duplicated Yeast Histones Suggest Strong Selection on a Coadapted Macromolecular Complex

We find evidence for interlocus gene conversion in five duplicated histone genes from six yeast species. The sequences of these duplicated genes, surviving from the ancient genome duplication, show phylogenetic patterns inconsistent with the well-resolved orthology relationships inferred from a likelihood model of gene loss after the genome duplication. Instead, these paralogous genes are more closely related to each other than any is to its nearest ortholog. In addition to simulations supporting gene conversion, we also present evidence for elevated rates of radical amino acid substitutions along the branches implicated in the conversion events. As these patterns are similar to those seen in ribosomal proteins that have undergone gene conversion, we speculate that in cases where duplicated genes code for proteins that are a part of tightly interacting complexes, selection may favor the fixation of gene conversion events in order to maintain high protein identities between duplicated copies.

Identification of 2R-ohnologue gene families displaying the same mutation-load skew in multiple cancers

The complexity of signalling pathways was boosted at the origin of the vertebrates, when two rounds of whole genome duplication (2R-WGD) occurred. Those genes and proteins that have survived from the 2R-WGD-termed 2R-ohnologues-belong to families of two to four members, and are enriched in signalling components relevant to cancer. Here, we find that while only approximately 30% of human transcript-coding genes are 2R-ohnologues, they carry 42-60% of the gene mutations in 30 different cancer types. Across a subset of cancer datasets, including melanoma, breast, lung adenocarcinoma, liver and medulloblastoma, we identified 673 2R-ohnologue families in which one gene carries mutations at multiple positions, while sister genes in the same family are relatively mutation free. Strikingly, in 315 of the 322 2R-ohnologue families displaying such a skew in multiple cancers, the same gene carries the heaviest mutation load in each cancer, and usually the second-ranked gene is also the same in each cancer. Our findings inspire the hypothesis that in certain cancers, heterogeneous combinations of genetic changes impair parts of the 2R-WGD signalling networks and force information flow through a limited set of oncogenic pathways in which specific non-mutated 2R-ohnologues serve as effectors. The non-mutated 2R-ohnologues are therefore potential therapeutic targets. These include proteins linked to growth factor signalling, neurotransmission and ion channels.

Evolution of signal multiplexing by 14-3-3-binding 2R-ohnologue protein families in the vertebrates

14-3-3 proteins regulate cellular responses to stimuli by docking onto pairs of phosphorylated residues on target proteins. The present study shows that the human 14-3-3-binding phosphoproteome is highly enriched in 2R-ohnologues, which are proteins in families of two to four members that were generated by two rounds of whole genome duplication at the origin of the vertebrates. We identify 2R-ohnologue families whose members share a ‘lynchpin’, defined as a 14-3-3-binding phosphosite that is conserved across members of a given family, and aligns with a Ser/Thr residue in pro-orthologues from the invertebrate chordates. For example, the human receptor expression enhancing protein (REEP) 1–4 family has the commonest type of lynchpin motif in current datasets, with a phosphorylatable serine in the –2 position relative to the 14-3-3-binding phosphosite. In contrast, the second 14-3-3-binding sites of REEPs 1–4 differ and are phosphorylated by different kinases, and hence the REEPs display different affinities for 14-3-3 dimers. We suggest a conceptual model for intracellular regulation involving protein families whose evolution into signal multiplexing systems was facilitated by 14-3-3 dimer binding to lynchpins, which gave freedom for other regulatory sites to evolve. While increased signalling complexity was needed for vertebrate life, these systems also generate vulnerability to genetic disorders.

Proteome-wide prediction of self-interacting proteins based on multiple properties

Self-interacting proteins, whose two or more copies can interact with each other, play important roles in cellular functions and the evolution of protein interaction networks (PINs). Knowing whether a protein can self-interact can contribute to and sometimes is crucial for the elucidation of its functions. Previous related research has mainly focused on the structures and functions of specific self-interacting proteins, whereas knowledge on their overall properties is limited. Meanwhile, the two current most common high throughput protein interaction assays have limited ability to detect self-interactions because of biological artifacts and design limitations, whereas the bioinformatic prediction method of self-interacting proteins is lacking. This study aims to systematically study and predict self-interacting proteins from an overall perspective. We find that compared with other proteins the self-interacting proteins in the structural aspect contain more domains; in the evolutionary aspect they tend to be conserved and ancient; in the functional aspect they are significantly enriched with enzyme genes, housekeeping genes, and drug targets, and in the topological aspect tend to occupy important positions in PINs. Furthermore, based on these features, after feature selection, we use logistic regression to integrate six representative features, including Gene Ontology term, domain, paralogous interactor, enzyme, model organism self-interacting protein, and betweenness centrality in the PIN, to develop a proteome-wide prediction model of self-interacting proteins. Using 5-fold cross-validation and an independent test, this model shows good performance. Finally, the prediction model is developed into a user-friendly web service SLIPPER (SeLf-Interacting Protein PrEdictoR). Users may submit a list of proteins, and then SLIPPER will return the probability_scores measuring their possibility to be self-interacting proteins and various related annotation information. This work helps us understand the role self-interacting proteins play in cellular functions from an overall perspective, and the constructed prediction model may contribute to the high throughput finding of self-interacting proteins and provide clues for elucidating their functions.

A conserved mammalian protein interaction network

Physical interactions between proteins mediate a variety of biological functions, including signal transduction, physical structuring of the cell and regulation. While extensive catalogs of such interactions are known from model organisms, their evolutionary histories are difficult to study given the lack of interaction data from phylogenetic outgroups. Using phylogenomic approaches, we infer a upper bound on the time of origin for a large set of human protein-protein interactions, showing that most such interactions appear relatively ancient, dating no later than the radiation of placental mammals. By analyzing paired alignments of orthologous and putatively interacting protein-coding genes from eight mammals, we find evidence for weak but significant co-evolution, as measured by relative selective constraint, between pairs of genes with interacting proteins. However, we find no strong evidence for shared instances of directional selection within an interacting pair. Finally, we use a network approach to show that the distribution of selective constraint across the protein interaction network is non-random, with a clear tendency for interacting proteins to share similar selective constraints. Collectively, the results suggest that, on the whole, protein interactions in mammals are under selective constraint, presumably due to their functional roles.

Patterns of indirect protein interactions suggest a spatial organization to metabolism

It has long been believed that cells organize their cytoplasm so as to efficiently channel metabolites between sequential enzymes. This metabolic channeling has the potential to yield higher metabolic fluxes as well as better regulatory control over metabolism. One mechanism for achieving such channeling is to ensure that sequential enzymes in a pathway are physically close to each other in the cell. We present evidence that indirect protein interactions between related enzymes represent a global mechanism for achieving metabolic channeling; the intuition being that protein interactions between enzymes and non-enzymatic mediator proteins are a powerful means of physically associating enzymes in a modular fashion. By analyzing the metabolic and protein-protein interactions networks of Escherichia coli, yeast and humans, we are able to show that all three species have many more indirect protein interactions linking enzymes that share metabolites than would be expected by chance. Moreover, these interactions are distributed non-randomly in the metabolic network. Our analyses in yeast and E. coli show that reactions possessing such interactions also show higher flux than do those lacking them. On the basis of these observations, we suggest that an important role of protein interactions with mediator proteins is to contribute to the spatial organization of the cell. This hypothesis is supported by the fact that these mediator proteins are also enriched with annotations related to signal transduction, a system where scaffolding proteins are known to limit cross-talk by controlling spatial localization.