Systematic Functional Annotation and Visualization of Biological Networks

High-throughput fitness experiments reveal specific vulnerabilities of human-adapted Salmonella during stress and infection

Salmonella enterica is comprised of genetically distinct 'serovars' that together provide an intriguing model for exploring the genetic basis of pathogen evolution. Although the genomes of numerous Salmonella isolates with broad variations in host range and human disease manifestations have been sequenced, the functional links between genetic and phenotypic differences among these serovars remain poorly understood. Here, we conduct high-throughput functional genomics on both generalist (Typhimurium) and human-restricted (Typhi and Paratyphi A) Salmonella at unprecedented scale in the study of this enteric pathogen. Using a comprehensive systems biology approach, we identify gene networks with serovar-specific fitness effects across 25 host-associated stresses encountered at key stages of human infection. By experimentally perturbing these networks, we characterize previously undescribed pseudogenes in human-adapted Salmonella. Overall, this work highlights specific vulnerabilities encoded within human-restricted Salmonella that are linked to the degradation of their genomes, shedding light into the evolution of this enteric pathogen.

Genome-scale analysis of interactions between genetic perturbations and natural variation

Interactions between genetic perturbations and segregating loci can cause perturbations to show different phenotypic effects across genetically distinct individuals. To study these interactions on a genome scale in many individuals, we used combinatorial DNA barcode sequencing to measure the fitness effects of 8046 CRISPRi perturbations targeting 1721 distinct genes in 169 yeast cross progeny (or segregants). We identified 460 genes whose perturbation has different effects across segregants. Several factors caused perturbations to show variable effects, including baseline segregant fitness, the mean effect of a perturbation across segregants, and interacting loci. We mapped 234 interacting loci and found four hub loci that interact with many different perturbations. Perturbations that interact with a given hub exhibit similar epistatic relationships with the hub and show enrichment for cellular processes that may mediate these interactions. These results suggest that an individual's response to perturbations is shaped by a network of perturbation-locus interactions that cannot be measured by approaches that examine perturbations or natural variation alone.

Elimination of virus-like particles reduces protein aggregation and extends replicative lifespan in Saccharomyces cerevisiae

A major consequence of aging and stress, in yeast to humans, is an increased accumulation of protein aggregates at distinct sites within the cells. Using genetic screens, immunoelectron microscopy, and three-dimensional modeling in our efforts to elucidate the importance of aggregate annexation, we found that most aggregates in yeast accumulate near the surface of mitochondria. Further, we show that virus-like particles (VLPs), which are part of the retrotransposition cycle of Ty elements, are markedly enriched in these sites of protein aggregation. RNA interference-mediated silencing of Ty expression perturbed aggregate sequestration to mitochondria, reduced overall protein aggregation, mitigated toxicity of a Huntington's disease model, and expanded the replicative lifespan of yeast in a partially Hsp104-dependent manner. The results are in line with recent data demonstrating that VLPs might act as aging factors in mammals, including humans, and extend these findings by linking VLPs to a toxic accumulation of protein aggregates and raising the possibility that they might negatively influence neurological disease progression.

High plasticity of ribosomal DNA organization in budding yeast

In eukaryotic genomes, rDNA generally resides as a highly repetitive and dynamic structure, making it difficult to study. Here, a synthetic rDNA array on chromosome III in budding yeast was constructed to serve as the sole source of rRNA. Utilizing the loxPsym site within each rDNA repeat and the Cre recombinase, we were able to reduce the copy number to as few as eight copies. Additionally, we constructed strains with two or three rDNA arrays and found that the presence of multiple arrays did not affect the formation of a single nucleolus. Although alteration of the position and number of rDNA arrays did impact the three-dimensional genome structure, the additional rDNA arrays had no deleterious influence on cell growth or transcriptomes. Overall, this study sheds light on the high plasticity of rDNA organization and opens up opportunities for future rDNA engineering.

Combinatorial prediction of therapeutic perturbations using causally-inspired neural networks

As an alternative to target-driven drug discovery, phenotype-driven approaches identify compounds that counteract the overall disease effects by analyzing phenotypic signatures. Our study introduces a novel approach to this field, aiming to expand the search space for new therapeutic agents. We introduce PDGrapher, a causally-inspired graph neural network model designed to predict arbitrary perturbagens - sets of therapeutic targets - capable of reversing disease effects. Unlike existing methods that learn responses to perturbations, PDGrapher solves the inverse problem, which is to infer the perturbagens necessary to achieve a specific response - i.e., directly predicting perturbagens by learning which perturbations elicit a desired response. Experiments across eight datasets of genetic and chemical perturbations show that PDGrapher successfully predicted effective perturbagens in up to 9% additional test samples and ranked therapeutic targets up to 35% higher than competing methods. A key innovation of PDGrapher is its direct prediction capability, which contrasts with the indirect, computationally intensive models traditionally used in phenotypedriven drug discovery that only predict changes in phenotypes due to perturbations. The direct approach enables PDGrapher to train up to 30 times faster, representing a significant leap in efficiency. Our results suggest that PDGrapher can advance phenotype-driven drug discovery, offering a fast and comprehensive approach to identifying therapeutically useful perturbations.

Genetic interactions of schizophrenia using gene-based statistical epistasis exclusively identify nervous system-related pathways and key hub genes

Background: The relationship between genotype and phenotype is governed by numerous genetic interactions (GIs), and the mapping of GI networks is of interest for two main reasons: 1) By modelling biological robustness, GIs provide a powerful opportunity to infer compensatory biological mechanisms via the identification of functional relationships between genes, which is of interest for biological discovery and translational research. Biological systems have evolved to compensate for genetic (i.e., variations and mutations) and environmental (i.e., drug efficacy) perturbations by exploiting compensatory relationships between genes, pathways and biological processes; 2) GI facilitates the identification of the direction (alleviating or aggravating interactions) and magnitude of epistatic interactions that influence the phenotypic outcome. The generation of GIs for human diseases is impossible using experimental biology approaches such as systematic deletion analysis. Moreover, the generation of disease-specific GIs has never been undertaken in humans. Methods: We used our Indian schizophrenia case-control (case-816, controls-900) genetic dataset to implement the workflow. Standard GWAS sample quality control procedure was followed. We used the imputed genetic data to increase the SNP coverage to analyse epistatic interactions across the genome comprehensively. Using the odds ratio (OR), we identified the GIs that increase or decrease the risk of a disease phenotype. The SNP-based epistatic results were transformed into gene-based epistatic results. Results: We have developed a novel approach by conducting gene-based statistical epistatic analysis using an Indian schizophrenia case-control genetic dataset and transforming these results to infer GIs that increase the risk of schizophrenia. There were ∼9.5 million GIs with a p-value ≤ 1 × 10-5. Approximately 4.8 million GIs showed an increased risk (OR > 1.0), while ∼4.75 million GIs had a decreased risk (OR <1.0) for schizophrenia. Conclusion: Unlike model organisms, this approach is specifically viable in humans due to the availability of abundant disease-specific genome-wide genotype datasets. The study exclusively identified brain/nervous system-related processes, affirming the findings. This computational approach fills a critical gap by generating practically non-existent heritable disease-specific human GIs from human genetic data. These novel datasets can train innovative deep-learning models, potentially surpassing the limitations of conventional GWAS.

Truncating the spliceosomal 'rope protein' Prp45 results in Htz1 dependent phenotypes

Spliceosome assembly contributes an important but incompletely understood aspect of splicing regulation. Prp45 is a yeast splicing factor which runs as an extended fold through the spliceosome, and which may be important for bringing its components together. We performed a whole genome analysis of the genetic interaction network of the truncated allele of PRP45 (prp45(1-169)) using synthetic genetic array technology and found chromatin remodellers and modifiers as an enriched category. In agreement with related studies, H2A.Z-encoding HTZ1, and the components of SWR1, INO80, and SAGA complexes represented prominent interactors, with htz1 conferring the strongest growth defect. Because the truncation of Prp45 disproportionately affected low copy number transcripts of intron-containing genes, we prepared strains carrying intronless versions of SRB2, VPS75, or HRB1, the most affected cases with transcription-related function. Intron removal from SRB2, but not from the other genes, partly repaired some but not all the growth phenotypes identified in the genetic screen. The interaction of prp45(1-169) and htz1Δ was detectable even in cells with SRB2 intron deleted (srb2Δi). The less truncated variant, prp45(1-330), had a synthetic growth defect with htz1Δ at 16°C, which also persisted in the srb2Δi background. Moreover, htz1Δ enhanced prp45(1-330) dependent pre-mRNA hyper-accumulation of both high and low efficiency splicers, genes ECM33 and COF1, respectively. We conclude that while the expression defects of low expression intron-containing genes contribute to the genetic interactome of prp45(1-169), the genetic interactions between prp45 and htz1 alleles demonstrate the sensitivity of spliceosome assembly, delayed in prp45(1-169), to the chromatin environment.

Building a eukaryotic chromosome arm by de novo design and synthesis

The genome of an organism is inherited from its ancestor and continues to evolve over time, however, the extent to which the current version could be altered remains unknown. To probe the genome plasticity of Saccharomyces cerevisiae, here we replace the native left arm of chromosome XII (chrXIIL) with a linear artificial chromosome harboring small sets of reconstructed genes. We find that as few as 12 genes are sufficient for cell viability, whereas 25 genes are required to recover the partial fitness defects observed in the 12-gene strain. Next, we demonstrate that these genes can be reconstructed individually using synthetic regulatory sequences and recoded open-reading frames with a "one-amino-acid-one-codon" strategy to remain functional. Finally, a synthetic neochromsome with the reconstructed genes is assembled which could substitute chrXIIL for viability. Together, our work not only highlights the high plasticity of yeast genome, but also illustrates the possibility of making functional eukaryotic chromosomes from entirely artificial sequences.

The regulatory landscape of the yeast phosphoproteome

The cellular ability to react to environmental fluctuations depends on signaling networks that are controlled by the dynamic activities of kinases and phosphatases. Here, to gain insight into these stress-responsive phosphorylation networks, we generated a quantitative mass spectrometry-based atlas of early phosphoproteomic responses in Saccharomyces cerevisiae exposed to 101 environmental and chemical perturbations. We report phosphosites on 59% of the yeast proteome, with 18% of the proteome harboring a phosphosite that is regulated within 5 min of stress exposure. We identify shared and perturbation-specific stress response programs, uncover loss of phosphorylation as an integral early event, and dissect the interconnected regulatory landscape of kinase-substrate networks, as we exemplify with target of rapamycin signaling. We further reveal functional organization principles of the stress-responsive phosphoproteome based on phosphorylation site motifs, kinase activities, subcellular localizations, shared functions and pathway intersections. This information-rich map of 25,000 regulated phosphosites advances our understanding of signaling networks.

The yeast guanine nucleotide exchange factor Sec7 is a bottleneck in spatial protein quality control and detoxifies neurological disease proteins

ER-to-Golgi trafficking partakes in the sorting of misfolded cytoplasmic proteins to reduce their cytological toxicity. We show here that yeast Sec7, a protein involved in proliferation of the Golgi, is part of this pathway and participates in an Hsp70-dependent formation of insoluble protein deposits (IPOD). Sec7 associates with the disaggregase Hsp104 during a mild heat shock and increases the rate of Hsp104 diffusion in an Hsp70-dependent manner when overproduced. Sec7 overproduction increased formation of IPODs from smaller aggregates and mitigated the toxicity of Huntingtin exon-1 upon heat stress while Sec7 depletion increased sensitivity to aẞ42 of the Alzheimer's disease and α-synuclein of the Parkinson's disease, suggesting a role of Sec7 in mitigating proteotoxicity.

Global analysis of the yeast knockout phenome

Genome-wide phenotypic screens in the budding yeast Saccharomyces cerevisiae, enabled by its knockout collection, have produced the largest, richest, and most systematic phenotypic description of any organism. However, integrative analyses of this rich data source have been virtually impossible because of the lack of a central data repository and consistent metadata annotations. Here, we describe the aggregation, harmonization, and analysis of ~14,500 yeast knockout screens, which we call Yeast Phenome. Using this unique dataset, we characterized two unknown genes (YHR045W and YGL117W) and showed that tryptophan starvation is a by-product of many chemical treatments. Furthermore, we uncovered an exponential relationship between phenotypic similarity and intergenic distance, which suggests that gene positions in both yeast and human genomes are optimized for function.

A global genetic interaction network by single-cell imaging and machine learning

Cellular and organismal phenotypes are controlled by complex gene regulatory networks. However, reference maps of gene function are still scarce across different organisms. Here, we generated synthetic genetic interaction and cell morphology profiles of more than 6,800 genes in cultured Drosophila cells. The resulting map of genetic interactions was used for machine learning-based gene function discovery, assigning functions to genes in 47 modules. Furthermore, we devised Cytoclass as a method to dissect genetic interactions for discrete cell states at the single-cell resolution. This approach identified an interaction of Cdk2 and the Cop9 signalosome complex, triggering senescence-associated secretory phenotypes and immunogenic conversion in hemocytic cells. Together, our data constitute a genome-scale resource of functional gene profiles to uncover the mechanisms underlying genetic interactions and their plasticity at the single-cell level.

Next-generation large-scale binary protein interaction network for Drosophila melanogaster

Generating reference maps of interactome networks illuminates genetic studies by providing a protein-centric approach to finding new components of existing pathways, complexes, and processes. We apply state-of-the-art methods to identify binary protein-protein interactions (PPIs) for Drosophila melanogaster. Four all-by-all yeast two-hybrid (Y2H) screens of > 10,000 Drosophila proteins result in the 'FlyBi' dataset of 8723 PPIs among 2939 proteins. Testing subsets of data from FlyBi and previous PPI studies using an orthogonal assay allows for normalization of data quality; subsequent integration of FlyBi and previous data results in an expanded binary Drosophila reference interaction network, DroRI, comprising 17,232 interactions among 6511 proteins. We use FlyBi data to generate an autophagy network, then validate in vivo using autophagy-related assays. The deformed wings (dwg) gene encodes a protein that is both a regulator and a target of autophagy. Altogether, these resources provide a foundation for building new hypotheses regarding protein networks and function.

Characterization of proteome-size scaling by integrative omics reveals mechanisms of proliferation control in cancer

Almost all living cells maintain size uniformity through successive divisions. Proteins that over and underscale with size can act as rheostats, which regulate cell cycle progression. Using a multiomic strategy, we leveraged the heterogeneity of melanoma cell lines to identify peptides, transcripts, and phosphorylation events that differentially scale with cell size. Subscaling proteins are enriched in regulators of the DNA damage response and cell cycle progression, whereas super-scaling proteins included regulators of the cytoskeleton, extracellular matrix, and inflammatory response. Mathematical modeling suggested that decoupling growth and proliferative signaling may facilitate cell cycle entry over senescence in large cells when mitogenic signaling is decreased. Regression analysis reveals that up-regulation of TP53 or CDKN1A/p21CIP1 is characteristic of proliferative cancer cells with senescent-like sizes/proteomes. This study provides one of the first demonstrations of size-scaling phenomena in cancer and how morphology influences the chemistry of the cell.

neuromaps: structural and functional interpretation of brain maps

Imaging technologies are increasingly used to generate high-resolution reference maps of brain structure and function. Comparing experimentally generated maps to these reference maps facilitates cross-disciplinary scientific discovery. Although recent data sharing initiatives increase the accessibility of brain maps, data are often shared in disparate coordinate systems, precluding systematic and accurate comparisons. Here we introduce neuromaps, a toolbox for accessing, transforming and analyzing structural and functional brain annotations. We implement functionalities for generating high-quality transformations between four standard coordinate systems. The toolbox includes curated reference maps and biological ontologies of the human brain, such as molecular, microstructural, electrophysiological, developmental and functional ontologies. Robust quantitative assessment of map-to-map similarity is enabled via a suite of spatial autocorrelation-preserving null models. neuromaps combines open-access data with transparent functionality for standardizing and comparing brain maps, providing a systematic workflow for comprehensive structural and functional annotation enrichment analysis of the human brain.

A network view of human immune system and virus-human interaction

The immune system is highly networked and complex, which is continuously changing as encountering old and new pathogens. However, reductionism-based researches do not give a systematic understanding of the molecular mechanism of the immune response and viral pathogenesis. Here, we present HUMPPI-2022, a high-quality human protein-protein interaction (PPI) network, containing > 11,000 protein-coding genes with > 78,000 interactions. The network topology and functional characteristics analyses of the immune-related genes (IRGs) reveal that IRGs are mostly located in the center of the network and link genes of diverse biological processes, which may reflect the gene pleiotropy phenomenon. Moreover, the virus-human interactions reveal that pan-viral targets are mostly hubs, located in the center of the network and enriched in fundamental biological processes, but not for coronavirus. Finally, gene age effect was analyzed from the view of the host network for IRGs and virally-targeted genes (VTGs) during evolution, with IRGs gradually became hubs and integrated into host network through bridging functionally differentiated modules. Briefly, HUMPPI-2022 serves as a valuable resource for gaining a better understanding of the composition and evolution of human immune system, as well as the pathogenesis of viruses.

Repression of essential cell cycle genes increases cellular fitness

A network of transcription factors (TFs) coordinates transcription with cell cycle events in eukaryotes. Most TFs in the network are phosphorylated by cyclin-dependent kinase (CDK), which limits their activities during the cell cycle. Here, we investigate the physiological consequences of disrupting CDK regulation of the paralogous repressors Yhp1 and Yox1 in yeast. Blocking Yhp1/Yox1 phosphorylation increases their levels and decreases expression of essential cell cycle regulatory genes which, unexpectedly, increases cellular fitness in optimal growth conditions. Using synthetic genetic interaction screens, we find that Yhp1/Yox1 mutations improve the fitness of mutants with mitotic defects, including condensin mutants. Blocking Yhp1/Yox1 phosphorylation simultaneously accelerates the G1/S transition and delays mitotic exit, without decreasing proliferation rate. This mitotic delay partially reverses the chromosome segregation defect of condensin mutants, potentially explaining their increased fitness when combined with Yhp1/Yox1 phosphomutants. These findings reveal how altering expression of cell cycle genes leads to a redistribution of cell cycle timing and confers a fitness advantage to cells.

COVID-19 severity is associated with population-level gut microbiome variations

The human gut microbiome interacts with many diseases, with recent small studies suggesting a link with COVID-19 severity. Exploring this association at the population-level may provide novel insights and help to explain differences in COVID-19 severity between countries. We explore whether there is an association between the gut microbiome of people within different countries and the severity of COVID-19, measured as hospitalisation rate. We use data from the large (n = 3,055) open-access gut microbiome repository curatedMetagenomicData, as well as demographic and country-level metadata. Twelve countries were placed into two groups (high/low) according to COVID-19 hospitalisation rate before December 2020 (ourworldindata.com). We use an unsupervised machine learning method, Topological Data Analysis (TDA). This method analyses both the local geometry and global topology of a high-dimensional dataset, making it particularly suitable for population-level microbiome data. We report an association of distinct baseline population-level gut microbiome signatures with COVID-19 severity. This was found both with a PERMANOVA, as well as with TDA. Specifically, it suggests an association of anti-inflammatory bacteria, including Bifidobacteria species and Eubacterium rectale, with lower severity, and pro-inflammatory bacteria such as Prevotella copri with higher severity. This study also reports a significant impact of country-level confounders, specifically of the proportion of over 70-year-olds in the population, GDP, and human development index. Further interventional studies should examine whether these relationships are causal, as well as considering the contribution of other variables such as genetics, lifestyle, policy, and healthcare system. The results of this study support the value of a population-level association design in microbiome research in general and for the microbiome-COVID-19 relationship, in particular. Finally, this research underscores the potential of TDA for microbiome studies, and in identifying novel associations.

Genetic background and mistranslation frequency determine the impact of mistranslating tRNASerUGG

Transfer RNA variants increase the frequency of mistranslation, the misincorporation of an amino acid not specified by the “standard” genetic code, to frequencies approaching 10% in yeast and bacteria. Cells cope with these variants by having multiple copies of each tRNA isodecoder and through pathways that deal with proteotoxic stress. In this study, we define the genetic interactions of the gene encoding tRNA^Ser_UGG,G26A, which mistranslates serine at proline codons. Using a collection of yeast temperature-sensitive alleles, we identify negative synthetic genetic interactions between the mistranslating tRNA and 109 alleles representing 91 genes, with nearly half of the genes having roles in RNA processing or protein folding and turnover. By regulating tRNA expression, we then compare the strength of the negative genetic interaction for a subset of identified alleles under differing amounts of mistranslation. The frequency of mistranslation correlated with the impact on cell growth for all strains analyzed; however, there were notable differences in the extent of the synthetic interaction at different frequencies of mistranslation depending on the genetic background. For many of the strains, the extent of the negative interaction with tRNA^Ser_UGG,G26A was proportional to the frequency of mistranslation or only observed at intermediate or high frequencies. For others, the synthetic interaction was approximately equivalent at all frequencies of mistranslation. As humans contain similar mistranslating tRNAs, these results are important when analyzing the impact of tRNA variants on disease, where both the individual’s genetic background and the expression of the mistranslating tRNA variant need to be considered.

Complex mutation profiles in mismatch repair and ribonucleotide reductase mutants reveal novel repair substrate specificity of MutS homolog (MSH) complexes

Determining mutation signatures is standard for understanding the etiology of human tumors and informing cancer treatment. Multiple determinants of DNA replication fidelity prevent mutagenesis that leads to carcinogenesis, including the regulation of free deoxyribonucleoside triphosphate pools by ribonucleotide reductase and repair of replication errors by the mismatch repair system. We identified genetic interactions between rnr1 alleles that skew and/or elevate deoxyribonucleoside triphosphate levels and mismatch repair gene deletions. These defects indicate that the rnr1 alleles lead to increased mutation loads that are normally acted upon by mismatch repair. We then utilized a targeted deep-sequencing approach to determine mutational profiles associated with mismatch repair pathway defects. By combining rnr1 and msh mutations to alter and/or increase deoxyribonucleoside triphosphate levels and alter the mutational load, we uncovered previously unreported specificities of Msh2-Msh3 and Msh2-Msh6. Msh2-Msh3 is uniquely able to direct the repair of G/C single-base deletions in GC runs, while Msh2-Msh6 specifically directs the repair of substitutions that occur at G/C dinucleotides. We also identified broader sequence contexts that influence variant profiles in different genetic backgrounds. Finally, we observed that the mutation profiles in double mutants were not necessarily an additive relationship of mutation profiles in single mutants. Our results have implications for interpreting mutation signatures from human tumors, particularly when mismatch repair is defective.

A genome-wide atlas of antibiotic susceptibility targets and pathways to tolerance

Detailed knowledge on how bacteria evade antibiotics and eventually develop resistance could open avenues for novel therapeutics and diagnostics. It is thereby key to develop a comprehensive genome-wide understanding of how bacteria process antibiotic stress, and how modulation of the involved processes affects their ability to overcome said stress. Here we undertake a comprehensive genetic analysis of how the human pathogen Streptococcus pneumoniae responds to 20 antibiotics. We build a genome-wide atlas of drug susceptibility determinants and generated a genetic interaction network that connects cellular processes and genes of unknown function, which we show can be used as therapeutic targets. Pathway analysis reveals a genome-wide atlas of cellular processes that can make a bacterium less susceptible, and often tolerant, in an antibiotic specific manner. Importantly, modulation of these processes confers fitness benefits during active infections under antibiotic selection. Moreover, screening of sequenced clinical isolates demonstrates that mutations in genes that decrease antibiotic sensitivity and increase tolerance readily evolve and are frequently associated with resistant strains, indicating such mutations could be harbingers for the emergence of antibiotic resistance.

Elastic network modeling of cellular networks unveils sensor and effector genes that control information flow

The high-level organization of the cell is embedded in indirect relationships that connect distinct cellular processes. Existing computational approaches for detecting indirect relationships between genes typically consist of propagating abstract information through network representations of the cell. However, the selection of genes to serve as the source of propagation is inherently biased by prior knowledge. Here, we sought to derive an unbiased view of the high-level organization of the cell by identifying the genes that propagate and receive information most effectively in the cell, and the indirect relationships between these genes. To this aim, we adapted a perturbation-response scanning strategy initially developed for identifying allosteric interactions within proteins. We deployed this strategy onto an elastic network model of the yeast genetic interaction profile similarity network. This network revealed a superior propensity for information propagation relative to simulated networks with similar topology. Perturbation-response scanning identified the major distributors and receivers of information in the network, named effector and sensor genes, respectively. Effectors formed dense clusters centrally integrated into the network, whereas sensors formed loosely connected antenna-shaped clusters and contained genes with previously characterized involvement in signal transduction. We propose that indirect relationships between effector and sensor clusters represent major paths of information flow between distinct cellular processes. Genetic similarity networks for fission yeast and human displayed similarly strong propensities for information propagation and clusters of effector and sensor genes, suggesting that the global architecture enabling indirect relationships is evolutionarily conserved across species. Our results demonstrate that elastic network modeling of cellular networks constitutes a promising strategy to probe the high-level organization and cooperativity in the cell.

High and stable ATP levels prevent aberrant intracellular protein aggregation in yeast

Adenosine triphosphate (ATP) at millimolar levels has recently been implicated in the solubilization of cellular proteins. However, the significance of this high ATP level under physiological conditions and the mechanisms that maintain ATP remain unclear. We herein demonstrated that AMP-activated protein kinase (AMPK) and adenylate kinase (ADK) cooperated to maintain cellular ATP levels regardless of glucose levels. Single-cell imaging of ATP-reduced yeast mutants revealed that ATP levels in these mutants underwent stochastic and transient depletion, which promoted the cytotoxic aggregation of endogenous proteins and pathogenic proteins, such as huntingtin and α-synuclein. Moreover, pharmacological elevations in ATP levels in an ATP-reduced mutant prevented the accumulation of α-synuclein aggregates and its cytotoxicity. The present study demonstrates that cellular ATP homeostasis ensures proteostasis and revealed that suppressing the high volatility of cellular ATP levels prevented cytotoxic protein aggregation, implying that AMPK and ADK are important factors that prevent proteinopathies, such as neurodegenerative diseases.

Cells use a chemical called adenosine triphosphate (ATP) as a controllable source of energy. Like a battery, each ATP molecule contains a specific amount of energy that can be released when needed. Cells just need enough ATP to survive, but most cells store a lot more than they need. It is unclear why cells keep so much ATP, or whether this excess ATP has any other purpose.

To answer these questions, Takaine et al. identified mutants of the yeast Saccharomyces cerevisiae that had low levels of ATP, and studied how these cells differ from normal yeast The results showed that, in S. cerevisiae cells with lower and variable levels of ATP, proteins stick together, forming clumps. Proteins are molecules that perform diverse roles, keeping cells alive. When they clump together, they stop working and can cause cells to die. Further experiments showed that reducing the levels of ATP just for a short time increased the rate at which proteins stick together.

Taken together, Takaine et al.’s results suggest that ATP plays a role in stopping proteins from sticking together, explaining why cells may store excess ATP, since it could aid survival.

Protein clumps, also called aggregates, are a key feature of various illnesses, including neurodegenerative diseases such as Alzheimer’s. Takaine et al. provide a possible cause for why proteins aggregate in these diseases, which may be worth further study.

Creeping yeast: a simple, cheap, and robust protocol for the identification of mating type in Saccharomyces cerevisiae

Saccharomyces cerevisiae is an exceptional genetic system, with genetic crosses facilitated by its ability to be maintained in haploid and diploid forms. Such crosses are straightforward if the mating type/ploidy of the strains is known. Several techniques can determine mating type (or ploidy), but all have limitations. Here, we validate a simple, cheap and robust method to identify S. cerevisiae mating types. When cells of opposite mating type are mixed in liquid media, they ‘creep’ up the culture vessel sides, a phenotype that can be easily detected visually. In contrast, mixtures of the same mating type or with a diploid simply settle out. The phenotype is observable for several days under a range of routine growth conditions and with different media/strains. Microscopy suggests that cell aggregation during mating is responsible for the phenotype. Yeast knockout collection analysis identified 107 genes required for the creeping phenotype, with these being enriched for mating-specific genes. Surprisingly, the RIM101 signaling pathway was strongly represented. We propose that RIM101 signaling regulates aggregation as part of a wider, previously unrecognized role in mating. The simplicity and robustness of this method make it ideal for routine verification of S. cerevisiae mating type, with future studies required to verify its molecular basis.

High-throughput functional characterization of protein phosphorylation sites in yeast

Phosphorylation is a critical post-translational modification involved in the regulation of almost all cellular processes. However, fewer than 5% of thousands of recently discovered phosphosites have been functionally annotated. In this study, we devised a chemical genetic approach to study the functional relevance of phosphosites in Saccharomyces cerevisiae. We generated 474 yeast strains with mutations in specific phosphosites that were screened for fitness in 102 conditions, along with a gene deletion library. Of these phosphosites, 42% exhibited growth phenotypes, suggesting that these are more likely functional. We inferred their function based on the similarity of their growth profiles with that of gene deletions and validated a subset by thermal proteome profiling and lipidomics. A high fraction exhibited phenotypes not seen in the corresponding gene deletion, suggestive of a gain-of-function effect. For phosphosites conserved in humans, the severity of the yeast phenotypes is indicative of their human functional relevance. This high-throughput approach allows for functionally characterizing individual phosphosites at scale.

Functional buffering via cell-specific gene expression promotes tissue homeostasis and cancer robustness

Functional buffering that ensures biological robustness is critical for maintaining tissue homeostasis, organismal survival, and evolution of novelty. However, the mechanism underlying functional buffering, particularly in multicellular organisms, remains largely elusive. Here, we proposed that functional buffering can be mediated via expression of buffering genes in specific cells and tissues, by which we named Cell-specific Expression-BUffering (CEBU). We developed an inference index (C-score) for CEBU by computing C-scores across 684 human cell lines using genome-wide CRISPR screens and transcriptomic RNA-seq. We report that C-score-identified putative buffering gene pairs are enriched for members of the same duplicated gene family, pathway, and protein complex. Furthermore, CEBU is especially prevalent in tissues of low regenerative capacity (e.g., bone and neuronal tissues) and is weakest in highly regenerative blood cells, linking functional buffering to tissue regeneration. Clinically, the buffering capacity enabled by CEBU can help predict patient survival for multiple cancers. Our results suggest CEBU as a potential buffering mechanism contributing to tissue homeostasis and cancer robustness in humans.

Network cartographs for interpretable visualizations

Networks offer an intuitive visual representation of complex systems. Important network characteristics can often be recognized by eye and, in turn, patterns that stand out visually often have a meaningful interpretation. In conventional network layout algorithms, however, the precise determinants of a node's position within a layout are difficult to decipher and to control. Here we propose an approach for directly encoding arbitrary structural or functional network characteristics into node positions. We introduce a series of two- and three-dimensional layouts, benchmark their efficiency for model networks, and demonstrate their power for elucidating structure-to-function relationships in large-scale biological networks.

Spatiotemporal localization of proteins in mycobacteria

Although prokaryotic organisms lack traditional organelles, they must still organize cellular structures in space and time, challenges that different species solve differently. To systematically define the subcellular architecture of mycobacteria, we perform high-throughput imaging of a library of fluorescently tagged proteins expressed in Mycobacterium smegmatis and develop a customized computational pipeline, MOMIA and GEMATRIA, to analyze these data. Our results establish a spatial organization network of over 700 conserved mycobacterial proteins and reveal a coherent localization pattern for many proteins of known function, including those in translation, energy metabolism, cell growth and division, as well as proteins of unknown function. Furthermore, our pipeline exploits morphologic proxies to enable a pseudo-temporal approximation of protein localization and identifies previously uncharacterized cell-cycle-dependent dynamics of essential mycobacterial proteins. Collectively, these data provide a systems perspective on the subcellular organization of mycobacteria and provide tools for the analysis of bacteria with non-standard growth characteristics.

Zhu et al. develop a two-stage image analysis pipeline, MOMIA and GEMATRIA, that efficiently models the spatial and temporal dynamics of over 700 conserved proteins in M. smegmatis. Through the analysis they report spatial constraints of mycobacterial ribosomes and membrane complexes and reconstruct temporal dynamics from still image data.

VID22 counteracts G-quadruplex-induced genome instability

Genome instability is a condition characterized by the accumulation of genetic alterations and is a hallmark of cancer cells. To uncover new genes and cellular pathways affecting endogenous DNA damage and genome integrity, we exploited a Synthetic Genetic Array (SGA)-based screen in yeast. Among the positive genes, we identified VID22, reported to be involved in DNA double-strand break repair. vid22Δ cells exhibit increased levels of endogenous DNA damage, chronic DNA damage response activation and accumulate DNA aberrations in sequences displaying high probabilities of forming G-quadruplexes (G4-DNA). If not resolved, these DNA secondary structures can block the progression of both DNA and RNA polymerases and correlate with chromosome fragile sites. Vid22 binds to and protects DNA at G4-containing regions both in vitro and in vivo. Loss of VID22 causes an increase in gross chromosomal rearrangement (GCR) events dependent on G-quadruplex forming sequences. Moreover, the absence of Vid22 causes defects in the correct maintenance of G4-DNA rich elements, such as telomeres and mtDNA, and hypersensitivity to the G4-stabilizing ligand TMPyP4. We thus propose that Vid22 is directly involved in genome integrity maintenance as a novel regulator of G4 metabolism.

Golgi membrane protein Erd1 Is essential for recycling a subset of Golgi glycosyltransferases

Protein glycosylation in the Golgi is a sequential process that requires proper distribution of transmembrane glycosyltransferase enzymes in the appropriate Golgi compartments. Some of the cytosolic machinery required for the steady-state localization of some Golgi enzymes are known but existing models do not explain how many of these enzymes are localized. Here, we uncover the role of an integral membrane protein in yeast, Erd1, as a key facilitator of Golgi glycosyltransferase recycling by directly interacting with both the Golgi enzymes and the cytosolic receptor, Vps74. Loss of Erd1 function results in mislocalization of Golgi enzymes to the vacuole/lysosome. We present evidence that Erd1 forms an integral part of the recycling machinery and ensures productive recycling of several early Golgi enzymes. Our work provides new insights on how the localization of Golgi glycosyltransferases is spatially and temporally regulated, and is finely tuned to the cues of Golgi maturation.

Topological Data Analysis Highlights Novel Geographical Signatures of the Human Gut Microbiome

Background: There is growing interest in the connection between the gut microbiome and human health and disease. Conventional approaches to analyse microbiome data typically entail dimensionality reduction and assume linearity of the observed relationships, however, the microbiome is a highly complex ecosystem marked by non-linear relationships. In this study, we use topological data analysis (TDA) to explore differences and similarities between the gut microbiome across several countries. Methods: We used curated adult microbiome data at the genus level from the GMrepo database. The dataset contains OTU and demographical data of over 4,400 samples from 19 studies, spanning 12 countries. We analysed the data with tmap, an integrative framework for TDA specifically designed for stratification and enrichment analysis of population-based gut microbiome datasets. Results: We find associations between specific microbial genera and groups of countries. Specifically, both the USA and UK were significantly co-enriched with the proinflammatory genera Lachnoclostridium and Ruminiclostridium, while France and New Zealand were co-enriched with other, butyrate-producing, taxa of the order Clostridiales. Conclusion: The TDA approach demonstrates the overlap and distinctions of microbiome composition between and within countries. This yields unique insights into complex associations in the dataset, a finding not possible with conventional approaches. It highlights the potential utility of TDA as a complementary tool in microbiome research, particularly for large population-scale datasets, and suggests further analysis on the effects of diet and other regionally varying factors.

Investigation of RNA metabolism through large-scale genetic interaction profiling in yeast

Gene deletion and gene expression alteration can lead to growth defects that are amplified or reduced when a second mutation is present in the same cells. We performed 154 genetic interaction mapping (GIM) screens with query mutants related with RNA metabolism and estimated the growth rates of about 700 000 double mutant Saccharomyces cerevisiae strains. The tested targets included the gene deletion collection and 900 strains in which essential genes were affected by mRNA destabilization (DAmP). To analyze the results, we developed RECAP, a strategy that validates genetic interaction profiles by comparison with gene co-citation frequency, and identified links between 1471 genes and 117 biological processes. In addition to these large-scale results, we validated both enhancement and suppression of slow growth measured for specific RNA-related pathways. Thus, negative genetic interactions identified a role for the OCA inositol polyphosphate hydrolase complex in mRNA translation initiation. By analysis of suppressors, we found that Puf4, a Pumilio family RNA binding protein, inhibits ribosomal protein Rpl9 function, by acting on a conserved UGUAcauUA motif located downstream the stop codon of the RPL9B mRNA. Altogether, the results and their analysis should represent a useful resource for discovery of gene function in yeast.

An Hsp90 co-chaperone links protein folding and degradation and is part of a conserved protein quality control

In this paper, we show that the essential Hsp90 co-chaperone Sgt1 is a member of a general protein quality control network that links folding and degradation through its participation in the degradation of misfolded proteins both in the cytosol and the endoplasmic reticulum (ER). Sgt1-dependent protein degradation acts in a parallel pathway to the ubiquitin ligase (E3) and ubiquitin chain elongase (E4), Hul5, and overproduction of Hul5 partly suppresses defects in cells with reduced Sgt1 activity. Upon proteostatic stress, Sgt1 accumulates transiently, in an Hsp90- and proteasome-dependent manner, with quality control sites (Q-bodies) of both yeast and human cells that co-localize with Vps13, a protein that creates organelle contact sites. Misfolding disease proteins, such as synphilin-1 involved in Parkinson's disease, are also sequestered to these compartments and require Sgt1 for their clearance.

Environmental robustness of the global yeast genetic interaction network

Phenotypes associated with genetic variants can be altered by interactions with other genetic variants (GxG), with the environment (GxE), or both (GxGxE). Yeast genetic interactions have been mapped on a global scale, but the environmental influence on the plasticity of genetic networks has not been examined systematically. To assess environmental rewiring of genetic networks, we examined 14 diverse conditions and scored 30,000 functionally representative yeast gene pairs for dynamic, differential interactions. Different conditions revealed novel differential interactions, which often uncovered functional connections between distantly related gene pairs. However, the majority of observed genetic interactions remained unchanged in different conditions, suggesting that the global yeast genetic interaction network is robust to environmental perturbation and captures the fundamental functional architecture of a eukaryotic cell.

Proteome-wide Prediction of Lysine Methylation Leads to Identification of H2BK43 Methylation and Outlines the Potential Methyllysine Proteome

Protein Lys methylation plays a critical role in numerous cellular processes, but it is challenging to identify Lys methylation in a systematic manner. Here we present an approach combining in silico prediction with targeted mass spectrometry (MS) to identify Lys methylation (Kme) sites at the proteome level. We develop MethylSight, a program that predicts Kme events solely on the physicochemical properties of residues surrounding the putative methylation sites, which then requires validation by targeted MS. Using this approach, we identify 70 new histone Kme marks with a 90% validation rate. H2BK43me2, which undergoes dynamic changes during stem cell differentiation, is found to be a substrate of KDM5b. Furthermore, MethylSight predicts that Lys methylation is a prevalent post-translational modification in the human proteome. Our work provides a useful resource for guiding systematic exploration of the role of Lys methylation in human health and disease.

τ-SGA: synthetic genetic array analysis for systematically screening and quantifying trigenic interactions in yeast

Systematic complex genetic interaction studies have provided insight into high-order functional redundancies and genetic network wiring of the cell. Here, we describe a method for screening and quantifying trigenic interactions from ordered arrays of yeast strains grown on agar plates as individual colonies. The protocol instructs users on the trigenic synthetic genetic array analysis technique, τ-SGA, for high-throughput screens. The steps describe construction of the double-mutant query strains and the corresponding single-mutant control query strains, which are screened in parallel in two replicates. The screening experimental set-up consists of sequential replica-pinning steps that enable automated mating, meiotic recombination and successive haploid selection steps for the generation of triple mutants, which are scored for colony size as a proxy for fitness, which enables the calculation of trigenic interactions. The procedure described here was used to conduct 422 trigenic interaction screens, which generated ~460,000 yeast triple mutants for trigenic interaction analysis. Users should be familiar with robotic equipment required for high-throughput genetic interaction screens and be proficient at the command line to execute the scoring pipeline. Large-scale screen computational analysis is achieved by using MATLAB pipelines that score raw colony size data to produce τ-SGA interaction scores. Additional recommendations are included for optimizing experimental design and analysis of smaller-scale trigenic interaction screens by using a web-based analysis system, SGAtools. This protocol provides a resource for those who would like to gain a deeper, more practical understanding of trigenic interaction screening and quantification methodology.

Construction of Functional Protein Networks Using Domain Profile Associations

Proteins are major functional molecules that physically and functionally interact to carry out cellular processes. The physical interactions are generally mediated by domain-level interactions. Thus, novel protein-protein interactions can be predicted using various computational methods based on domain-domain interactions, using resolved structures of protein complexes. Functional protein interactions can be inferred based on shared domains between proteins, since proteins involved in the same biological processes tend to harbor common domains. We recently developed a method of inferring functional interactions between proteins using associations between their domain compositions, which can be represented as domain profiles. Since the method requires only protein domain annotations, it can be easily applied to any species with a sequenced genome. Here, we describe in detail the method of generating domain profiles for proteins and measuring the association between them to infer functional interactions between proteins. We also demonstrate that domain profile association can be used to successfully construct a large-scale functional network of human proteins.

Trigenic Synthetic Genetic Array (τ-SGA) Technique for Complex Interaction Analysis

Complex genetic interactions occur when mutant alleles of multiple genes combine to elicit an unexpected phenotype, which could not be predicted given the expectation based on the combination of phenotypes associated with individual mutant alleles. Trigenic Synthetic Genetic Array (τ-SGA) methodology was developed for the systematic analysis of complex interactions involving combinations of three gene perturbations. With a series of replica pinning steps of the τ-SGA procedure, haploid triple mutants are constructed through automated mating and meiotic recombination. For example, a double-mutant query strain carrying two mutant alleles of interest, such as a deletion allele of a nonessential gene and a conditional temperature-sensitive allele of an essential gene, is crossed to an input array of yeast mutants, such as the diagnostic array set of ~1200 mutants, to generate an output array of triple mutants. The colony-size measurements of the resulting triple mutants are used to estimate cellular fitness and quantify trigenic interactions by incorporating corresponding single- and double-mutant fitness estimates. Trigenic interaction networks can be further analyzed for functional modules using various clustering and enrichment analysis tools. Complex genetic interactions are rich in functional information and provide insight into the genotype-to-phenotype relationship, genome size, and speciation.

Chemical-Genetic Interactions with the Proline Analog L-Azetidine-2-Carboxylic Acid in Saccharomyces cerevisiae

Non-proteinogenic amino acids, such as the proline analog L-azetidine-2-carboxylic acid (AZC), are detrimental to cells because they are mis-incorporated into proteins and lead to proteotoxic stress. Our goal was to identify genes that show chemical-genetic interactions with AZC in Saccharomyces cerevisiae and thus also potentially define the pathways cells use to cope with amino acid mis-incorporation. Screening the yeast deletion and temperature sensitive collections, we found 72 alleles with negative chemical-genetic interactions with AZC treatment and 12 alleles that suppress AZC toxicity. Many of the genes with negative chemical-genetic interactions are involved in protein quality control pathways through the proteasome. Genes involved in actin cytoskeleton organization and endocytosis also had negative chemical-genetic interactions with AZC. Related to this, the number of actin patches per cell increases upon AZC treatment. Many of the same cellular processes were identified to have interactions with proteotoxic stress caused by two other amino acid analogs, canavanine and thialysine, or a mistranslating tRNA variant that mis-incorporates serine at proline codons. Alleles that suppressed AZC-induced toxicity functioned through the amino acid sensing TOR pathway or controlled amino acid permeases required for AZC uptake. Further suggesting the potential of genetic changes to influence the cellular response to proteotoxic stress, overexpressing many of the genes that had a negative chemical-genetic interaction with AZC suppressed AZC toxicity.

Arrayed CRISPRi and quantitative imaging describe the morphotypic landscape of essential mycobacterial genes

Mycobacterium tuberculosis possesses a large number of genes of unknown or predicted function, undermining fundamental understanding of pathogenicity and drug susceptibility. To address this challenge, we developed a high-throughput functional genomics approach combining inducible CRISPR-interference and image-based analyses of morphological features and sub-cellular chromosomal localizations in the related non-pathogen, M. smegmatis. Applying automated imaging and analysis to 263 essential gene knockdown mutants in an arrayed library, we derive robust, quantitative descriptions of bacillary morphologies consequent on gene silencing. Leveraging statistical-learning, we demonstrate that functionally related genes cluster by morphotypic similarity and that this information can be used to inform investigations of gene function. Exploiting this observation, we infer the existence of a mycobacterial restriction-modification system, and identify filamentation as a defining mycobacterial response to histidine starvation. Our results support the application of large-scale image-based analyses for mycobacterial functional genomics, simultaneously establishing the utility of this approach for drug mechanism-of-action studies.

Syntaxin 5 Is Required for the Formation and Clearance of Protein Inclusions during Proteostatic Stress

Spatial sorting to discrete quality control sites in the cell is a process harnessing the toxicity of aberrant proteins. We show that the yeast t-snare phosphoprotein syntaxin5 (Sed5) acts as a key factor in mitigating proteotoxicity and the spatial deposition and clearance of IPOD (insoluble protein deposit) inclusions associates with the disaggregase Hsp104. Sed5 phosphorylation promotes dynamic movement of COPII-associated Hsp104 and boosts disaggregation by favoring anterograde ER-to-Golgi trafficking. Hsp104-associated aggregates co-localize with Sed5 as well as components of the ER, trans Golgi network, and endocytic vesicles, transiently during proteostatic stress, explaining mechanistically how misfolded and aggregated proteins formed at the vicinity of the ER can hitchhike toward vacuolar IPOD sites. Many inclusions become associated with mitochondria in a HOPS/vCLAMP-dependent manner and co-localize with Vps39 (HOPS/vCLAMP) and Vps13, which are proteins providing contacts between vacuole and mitochondria. Both Vps39 and Vps13 are required also for efficient Sed5-dependent clearance of aggregates.

Polo kinase recruitment via the constitutive centromere-associated network at the kinetochore elevates centromeric RNA

The kinetochore, a multi-protein complex assembled on centromeres, is essential to segregate chromosomes during cell division. Deficiencies in kinetochore function can lead to chromosomal instability and aneuploidy-a hallmark of cancer cells. Kinetochore function is controlled by recruitment of regulatory proteins, many of which have been documented, however their function often remains uncharacterized and many are yet to be identified. To identify candidates of kinetochore regulation we used a proteome-wide protein association strategy in budding yeast and detected many proteins that are involved in post-translational modifications such as kinases, phosphatases and histone modifiers. We focused on the Polo-like kinase, Cdc5, and interrogated which cellular components were sensitive to constitutive Cdc5 localization. The kinetochore is particularly sensitive to constitutive Cdc5 kinase activity. Targeting Cdc5 to different kinetochore subcomplexes produced diverse phenotypes, consistent with multiple distinct functions at the kinetochore. We show that targeting Cdc5 to the inner kinetochore, the constitutive centromere-associated network (CCAN), increases the levels of centromeric RNA via an SPT4 dependent mechanism.

Exploring whole-genome duplicate gene retention with complex genetic interaction analysis

Whole-genome duplication has played a central role in the genome evolution of many organisms, including the human genome. Most duplicated genes are eliminated, and factors that influence the retention of persisting duplicates remain poorly understood. We describe a systematic complex genetic interaction analysis with yeast paralogs derived from the whole-genome duplication event. Mapping of digenic interactions for a deletion mutant of each paralog, and of trigenic interactions for the double mutant, provides insight into their roles and a quantitative measure of their functional redundancy. Trigenic interaction analysis distinguishes two classes of paralogs: a more functionally divergent subset and another that retained more functional overlap. Gene feature analysis and modeling suggest that evolutionary trajectories of duplicated genes are dictated by combined functional and structural entanglement factors.

A Genome-Wide Screen for Genes Affecting Spontaneous Direct-Repeat Recombination in Saccharomyces cerevisiae

Homologous recombination is an important mechanism for genome integrity maintenance, and several homologous recombination genes are mutated in various cancers and cancer-prone syndromes. However, since in some cases homologous recombination can lead to mutagenic outcomes, this pathway must be tightly regulated, and mitotic hyper-recombination is a hallmark of genomic instability. We performed two screens in Saccharomyces cerevisiae for genes that, when deleted, cause hyper-recombination between direct repeats. One was performed with the classical patch and replica-plating method. The other was performed with a high-throughput replica-pinning technique that was designed to detect low-frequency events. This approach allowed us to validate the high-throughput replica-pinning methodology independently of the replicative aging context in which it was developed. Furthermore, by combining the two approaches, we were able to identify and validate 35 genes whose deletion causes elevated spontaneous direct-repeat recombination. Among these are mismatch repair genes, the Sgs1-Top3-Rmi1 complex, the RNase H2 complex, genes involved in the oxidative stress response, and a number of other DNA replication, repair and recombination genes. Since several of our hits are evolutionarily conserved, and repeated elements constitute a significant fraction of mammalian genomes, our work might be relevant for understanding genome integrity maintenance in humans.

CSN: unsupervised approach for inferring biological networks based on the genome alone

BACKGROUND

Most organisms cannot be cultivated, as they live in unique ecological conditions that cannot be mimicked in the lab. Understanding the functionality of those organisms' genes and their interactions by performing large-scale measurements of transcription levels, protein-protein interactions or metabolism, is extremely difficult and, in some cases, impossible. Thus, efficient algorithms for deciphering genome functionality based only on the genomic sequences with no other experimental measurements are needed.

RESULTS

In this study, we describe a novel algorithm that infers gene networks that we name Common Substring Network (CSN). The algorithm enables inferring novel regulatory relations among genes based only on the genomic sequence of a given organism and partial homolog/ortholog-based functional annotation. It can specifically infer the functional annotation of genes with unknown homology. This approach is based on the assumption that related genes, not necessarily homologs, tend to share sub-sequences, which may be related to common regulatory mechanisms, similar functionality of encoded proteins, common evolutionary history, and more. We demonstrate that CSNs, which are based on S. cerevisiae and E. coli genomes, have properties similar to 'traditional' biological networks inferred from experiments. Highly expressed genes tend to have higher degree nodes in the CSN, genes with similar protein functionality tend to be closer, and the CSN graph exhibits a power-law degree distribution. Also, we show how the CSN can be used for predicting gene interactions and functions.

CONCLUSIONS

The reported results suggest that 'silent' code inside the transcript can help to predict central features of biological networks and gene function. This approach can help researchers to understand the genome of novel microorganisms, analyze metagenomic data, and can help to decipher new gene functions.

AVAILABILITY

Our MATLAB implementation of CSN is available at https://www.cs.tau.ac.il/~tamirtul/CSN-Autogen.

Genetic interaction profiles of regulatory kinases differ between environmental conditions and cellular states

Cell growth and quiescence in eukaryotic cells is controlled by an evolutionarily conserved network of signaling pathways. Signal transduction networks operate to modulate a wide range of cellular processes and physiological properties when cells exit proliferative growth and initiate a quiescent state. How signaling networks function to respond to diverse signals that result in cell cycle exit and establishment of a quiescent state is poorly understood. Here, we studied the function of signaling pathways in quiescent cells using global genetic interaction mapping in the model eukaryotic cell, Saccharomyces cerevisiae (budding yeast). We performed pooled analysis of genotypes using molecular barcode sequencing (Bar-seq) to test the role of ~4,000 gene deletion mutants and ~12,000 pairwise interactions between all non-essential genes and the protein kinase genes TOR1, RIM15, and PHO85 in three different nutrient-restricted conditions in both proliferative and quiescent cells. We detect up to 10-fold more genetic interactions in quiescent cells than proliferative cells. We find that both individual gene effects and genetic interaction profiles vary depending on the specific pro-quiescence signal. The master regulator of quiescence, RIM15, shows distinct genetic interaction profiles in response to different starvation signals. However, vacuole-related functions show consistent genetic interactions with RIM15 in response to different starvation signals, suggesting that RIM15 integrates diverse signals to maintain protein homeostasis in quiescent cells. Our study expands genome-wide genetic interaction profiling to additional conditions, and phenotypes, and highlights the conditional dependence of epistasis.

SDN2GO: An Integrated Deep Learning Model for Protein Function Prediction

The assignment of function to proteins at a large scale is essential for understanding the molecular mechanism of life. However, only a very small percentage of the more than 179 million proteins in UniProtKB have Gene Ontology (GO) annotations supported by experimental evidence. In this paper, we proposed an integrated deep-learning-based classification model, named SDN2GO, to predict protein functions. SDN2GO applies convolutional neural networks to learn and extract features from sequences, protein domains, and known PPI networks, and then utilizes a weight classifier to integrate these features and achieve accurate predictions of GO terms. We constructed the training set and the independent test set according to the time-delayed principle of the Critical Assessment of Function Annotation (CAFA) and compared it with two highly competitive methods and the classic BLAST method on the independent test set. The results show that our method outperforms others on each sub-ontology of GO. We also investigated the performance of using protein domain information. We learned from the Natural Language Processing (NLP) to process domain information and pre-trained a deep learning sub-model to extract the comprehensive features of domains. The experimental results demonstrate that the domain features we obtained are much improved the performance of our model. Our deep learning models together with the data pre-processing scripts are publicly available as an open source software at https://github.com/Charrick/SDN2GO.

A reference map of the human binary protein interactome

Global insights into cellular organization and genome function require comprehensive understanding of the interactome networks that mediate genotype-phenotype relationships^1,2. Here, we present a human “all-by-all” reference interactome map of human binary protein interactions, or “HuRI”. With ~53,000 high-quality protein-protein interactions (PPIs), HuRI has approximately four times more such interactions than high-quality curated interactions from small-scale studies. Integrating HuRI with genome³, transcriptome⁴, and proteome⁵ data enables the study of cellular function within most physiological or pathological cellular contexts. We demonstrate the utility of HuRI in identifying specific subcellular roles of PPIs. Inferred tissue-specific networks reveal general principles for the formation of cellular context-specific functions and elucidate potential molecular mechanisms underlying tissue-specific phenotypes of Mendelian diseases. HuRI represents a systematic proteome-wide reference linking genomic variation to phenotypic outcomes.