Expanding TheCellVision.org: A Central Repository for Visualizing and Mining High-Content Cell Imaging Projects

We previously constructed TheCellVision.org, a central repository for visualizing and mining data from yeast high-content imaging projects. At its inception, TheCellVision.org housed two high-content screening (HCS) projects providing genome-scale protein abundance and localization information for the budding yeast Saccharomyces cerevisiae, as well as a comprehensive analysis of the morphology of its endocytic compartments upon systematic genetic perturbation of each yeast gene. Here, we report on the expansion of TheCellVision.org by the addition of two new HCS projects and the incorporation of new global functionalities. Specifically, TheCellVision.org now hosts images from the Cell Cycle Omics project, which describes genome-scale cell cycle-resolved dynamics in protein localization, protein concentration, gene expression, and translational efficiency in budding yeast. Moreover, it hosts PIFiA, a computational tool for image-based predictions of protein functional annotations. Across all its projects, TheCellVision.org now houses > 800,000 microscopy images along with computational tools for exploring both the images and their associated datasets. Together with the newly added global functionalities, which include the ability to query genes in any of the hosted projects using either yeast or human gene names, TheCellVision.org provides an expanding resource for single-cell eukaryotic biology.

The PRC2.1 Subcomplex Opposes G1 Progression through Regulation of CCND1 and CCND2

Progression through the G1 phase of the cell cycle is the most highly regulated step in cellular division. We employed a chemogenomics approach to discover novel cellular networks that regulate cell cycle progression. This approach uncovered functional clusters of genes that altered sensitivity of cells to inhibitors of the G1/S transition. Mutation of components of the Polycomb Repressor Complex 2 rescued growth inhibition caused by the CDK4/6 inhibitor palbociclib, but not to inhibitors of S phase or mitosis. In addition to its core catalytic subunits, mutation of the PRC2.1 accessory protein MTF2, but not the PRC2.2 protein JARID2, rendered cells resistant to palbociclib treatment. We found that PRC2.1 (MTF2), but not PRC2.2 (JARID2), was critical for promoting H3K27me3 deposition at CpG islands genome-wide and in promoters. This included the CpG islands in the promoter of the CDK4/6 cyclins CCND1 and CCND2, and loss of MTF2 lead to upregulation of both CCND1 and CCND2. Our results demonstrate a role for PRC2.1, but not PRC2.2, in promoting G1 progression.

Proteome-scale movements and compartment connectivity during the eukaryotic cell cycle

Cell cycle progression relies on coordinated changes in the composition and subcellular localization of the proteome. By applying two distinct convolutional neural networks on images of millions of live yeast cells, we resolved proteome-level dynamics in both concentration and localization during the cell cycle, with resolution of ∼20 subcellular localization classes. We show that a quarter of the proteome displays cell cycle periodicity, with proteins tending to be controlled either at the level of localization or concentration, but not both. Distinct levels of protein regulation are preferentially utilized for different aspects of the cell cycle, with changes in protein concentration being mostly involved in cell cycle control and changes in protein localization in the biophysical implementation of the cell cycle program. We present a resource for exploring global proteome dynamics during the cell cycle, which will aid in understanding a fundamental biological process at a systems level.

N-terminal acetylation shields proteins from degradation and promotes age-dependent motility and longevity

Most eukaryotic proteins are N-terminally acetylated, but the functional impact on a global scale has remained obscure. Using genome-wide CRISPR knockout screens in human cells, we reveal a strong genetic dependency between a major N-terminal acetyltransferase and specific ubiquitin ligases. Biochemical analyses uncover that both the ubiquitin ligase complex UBR4-KCMF1 and the acetyltransferase NatC recognize proteins bearing an unacetylated N-terminal methionine followed by a hydrophobic residue. NatC KO-induced protein degradation and phenotypes are reversed by UBR knockdown, demonstrating the central cellular role of this interplay. We reveal that loss of Drosophila NatC is associated with male sterility, reduced longevity, and age-dependent loss of motility due to developmental muscle defects. Remarkably, muscle-specific overexpression of UbcE2M, one of the proteins targeted for NatC KO-mediated degradation, suppresses defects of NatC deletion. In conclusion, NatC-mediated N-terminal acetylation acts as a protective mechanism against protein degradation, which is relevant for increased longevity and motility.

Reproducibility metrics for context-specific CRISPR screens

CRISPR screens are used extensively to systematically interrogate the phenotype-to-genotype problem. In contrast to early CRISPR screens, which defined core cell fitness genes, most current efforts now aim to identify context-specific phenotypes that differentiate a cell line, genetic background, or condition of interest, such as a drug treatment. While CRISPR-related technologies have shown great promise and a fast pace of innovation, a better understanding of standards and methods for quality assessment of CRISPR screen results is crucial to guide technology development and application. Specifically, many commonly used metrics for quantifying screen quality do not accurately measure the reproducibility of context-specific hits. We highlight the importance of reporting reproducibility statistics that directly relate to the purpose of the screen and suggest the use of metrics that are sensitive to context-specific signal. A record of this paper's transparent peer review process is included in the supplemental information.

K29-linked unanchored polyubiquitin chains disrupt ribosome biogenesis and direct ribosomal proteins to the Intranuclear Quality control compartment (INQ)

Ribosome assembly requires precise coordination between the production and assembly of ribosomal components. Mutations in ribosomal proteins that inhibit the assembly process or ribosome function are often associated with Ribosomopathies, some of which are linked to defects in proteostasis. In this study, we examine the interplay between several yeast proteostasis enzymes, including deubiquitylases (DUBs), Ubp2 and Ubp14, and E3 ligases, Ufd4 and Hul5, and we explore their roles in the regulation of the cellular levels of K29-linked unanchored polyubiquitin (polyUb) chains. Accumulating K29-linked unanchored polyUb chains associate with maturing ribosomes to disrupt their assembly, activate the Ribosome assembly stress response (RASTR), and lead to the sequestration of ribosomal proteins at the Intranuclear Quality control compartment (INQ). These findings reveal the physiological relevance of INQ and provide insights into mechanisms of cellular toxicity associated with Ribosomopathies.

Systematic analysis of bypass suppression of essential genes

Essential genes tend to be highly conserved across eukaryotes, but, in some cases, their critical roles can be bypassed through genetic rewiring. From a systematic analysis of 728 different essential yeast genes, we discovered that 124 (17%) were dispensable essential genes. Through whole-genome sequencing and detailed genetic analysis, we investigated the genetic interactions and genome alterations underlying bypass suppression. Dispensable essential genes often had paralogs, were enriched for genes encoding membrane-associated proteins, and were depleted for members of protein complexes. Functionally related genes frequently drove the bypass suppression interactions. These gene properties were predictive of essential gene dispensability and of specific suppressors among hundreds of genes on aneuploid chromosomes. Our findings identify yeast's core essential gene set and reveal that the properties of dispensable essential genes are conserved from yeast to human cells, correlating with human genes that display cell line-specific essentiality in the Cancer Dependency Map (DepMap) project.

Timer-based proteomic profiling of the ubiquitin-proteasome system reveals a substrate receptor of the GID ubiquitin ligase

Selective protein degradation by the ubiquitin-proteasome system (UPS) is involved in all cellular processes. However, the substrates and specificity of most UPS components are not well understood. Here we systematically characterized the UPS in Saccharomyces cerevisiae. Using fluorescent timers, we determined how loss of individual UPS components affects yeast proteome turnover, detecting phenotypes for 76% of E2, E3, and deubiquitinating enzymes. We exploit this dataset to gain insights into N-degron pathways, which target proteins carrying N-terminal degradation signals. We implicate Ubr1, an E3 of the Arg/N-degron pathway, in targeting mitochondrial proteins processed by the mitochondrial inner membrane protease. Moreover, we identify Ylr149c/Gid11 as a substrate receptor of the glucose-induced degradation-deficient (GID) complex, an E3 of the Pro/N-degron pathway. Our results suggest that Gid11 recognizes proteins with N-terminal threonines, expanding the specificity of the GID complex. This resource of potential substrates and relationships between UPS components enables exploring functions of selective protein degradation.

A genome-scale yeast library with inducible expression of individual genes

The ability to switch a gene from off to on and monitor dynamic changes provides a powerful approach for probing gene function and elucidating causal regulatory relationships. Here, we developed and characterized YETI (Yeast Estradiol strains with Titratable Induction), a collection in which > 5,600 yeast genes are engineered for transcriptional inducibility with single-gene precision at their native loci and without plasmids. Each strain contains SGA screening markers and a unique barcode, enabling high-throughput genetics. We characterized YETI using growth phenotyping and BAR-seq screens, and we used a YETI allele to identify the regulon of Rof1, showing that it acts to repress transcription. We observed that strains with inducible essential genes that have low native expression can often grow without inducer. Analysis of data from eukaryotic and prokaryotic systems shows that native expression is a variable that can bias promoter-perturbing screens, including CRISPRi. We engineered a second expression system, Z3 EB42, that gives lower expression than Z3 EV, a feature enabling conditional activation and repression of lowly expressed essential genes that grow without inducer in the YETI library.

Single-cell image analysis to explore cell-to-cell heterogeneity in isogenic populations

Single-cell image analysis provides a powerful approach for studying cell-to-cell heterogeneity, which is an important attribute of isogenic cell populations, from microbial cultures to individual cells in multicellular organisms. This phenotypic variability must be explained at a mechanistic level if biologists are to fully understand cellular function and address the genotype-to-phenotype relationship. Variability in single-cell phenotypes is obscured by bulk readouts or averaging of phenotypes from individual cells in a sample; thus, single-cell image analysis enables a higher resolution view of cellular function. Here, we consider examples of both small- and large-scale studies carried out with isogenic cell populations assessed by fluorescence microscopy, and we illustrate the advantages, challenges, and the promise of quantitative single-cell image analysis.

A method for benchmarking genetic screens reveals a predominant mitochondrial bias

We present FLEX (Functional evaluation of experimental perturbations), a pipeline that leverages several functional annotation resources to establish reference standards for benchmarking human genome-wide CRISPR screen data and methods for analyzing them. FLEX provides a quantitative measurement of the functional information captured by a given gene-pair dataset and a means to explore the diversity of functions captured by the input dataset. We apply FLEX to analyze data from the diverse cell line screens generated by the DepMap project. We identify a predominant mitochondria-associated signal within co-essentiality networks derived from these data and explore the basis of this signal. Our analysis and time-resolved CRISPR screens in a single cell line suggest that the variable phenotypes associated with mitochondria genes across cells may reflect screen dynamics and protein stability effects rather than genetic dependencies. We characterize this functional bias and demonstrate its relevance for interpreting differential hits in any CRISPR screening context. More generally, we demonstrate the utility of the FLEX pipeline for performing robust comparative evaluations of CRISPR screens or methods for processing them.

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.

Genetic profiling of protein burden and nuclear export overload

Overproduction (op) of proteins triggers cellular defects. One of the consequences of overproduction is the protein burden/cost, which is produced by an overloading of the protein synthesis process. However, the physiology of cells under a protein burden is not well characterized. We performed genetic profiling of protein burden by systematic analysis of genetic interactions between GFP-op, surveying both deletion and temperature-sensitive mutants in budding yeast. We also performed genetic profiling in cells with overproduction of triple-GFP (tGFP), and the nuclear export signal-containing tGFP (NES-tGFP). The mutants specifically interacted with GFP-op were suggestive of unexpected connections between actin-related processes like polarization and the protein burden, which was supported by morphological analysis. The tGFP-op interactions suggested that this protein probe overloads the proteasome, whereas those that interacted with NES-tGFP involved genes encoding components of the nuclear export process, providing a resource for further analysis of the protein burden and nuclear export overload.

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.

Systematic mapping of genetic interactions for de novo fatty acid synthesis identifies C12orf49 as a regulator of lipid metabolism

The de novo synthesis of fatty acids has emerged as a therapeutic target for various diseases, including cancer. Because cancer cells are intrinsically buffered to combat metabolic stress, it is important to understand how cells may adapt to the loss of de novo fatty acid biosynthesis. Here, we use pooled genome-wide CRISPR screens to systematically map genetic interactions (GIs) in human HAP1 cells carrying a loss-of-function mutation in fatty acid synthase (FASN), whose product catalyses the formation of long-chain fatty acids. FASN-mutant cells show a strong dependence on lipid uptake that is reflected in negative GIs with genes involved in the LDL receptor pathway, vesicle trafficking and protein glycosylation. Further support for these functional relationships is derived from additional GI screens in query cell lines deficient in other genes involved in lipid metabolism, including LDLR, SREBF1, SREBF2 and ACACA. Our GI profiles also identify a potential role for the previously uncharacterized gene C12orf49 (which we call LUR1) in regulation of exogenous lipid uptake through modulation of SREBF2 signalling in response to lipid starvation. Overall, our data highlight the genetic determinants underlying the cellular adaptation associated with loss of de novo fatty acid synthesis and demonstrate the power of systematic GI mapping for uncovering metabolic buffering mechanisms in human cells.

Systematic genetics and single-cell imaging reveal widespread morphological pleiotropy and cell-to-cell variability

Our ability to understand the genotype-to-phenotype relationship is hindered by the lack of detailed understanding of phenotypes at a single-cell level. To systematically assess cell-to-cell phenotypic variability, we combined automated yeast genetics, high-content screening and neural network-based image analysis of single cells, focussing on genes that influence the architecture of four subcellular compartments of the endocytic pathway as a model system. Our unbiased assessment of the morphology of these compartments-endocytic patch, actin patch, late endosome and vacuole-identified 17 distinct mutant phenotypes associated with ~1,600 genes (~30% of all yeast genes). Approximately half of these mutants exhibited multiple phenotypes, highlighting the extent of morphological pleiotropy. Quantitative analysis also revealed that incomplete penetrance was prevalent, with the majority of mutants exhibiting substantial variability in phenotype at the single-cell level. Our single-cell analysis enabled exploration of factors that contribute to incomplete penetrance and cellular heterogeneity, including replicative age, organelle inheritance and response to stress.

Genetic interaction networks in cancer cells

The genotype-to-phenotype relationship in health and disease is complex and influenced by both an individual's environment and their unique genome. Personal genetic variants can modulate gene function to generate a phenotype either through a single gene effect or through genetic interactions involving two or more genes. The relevance of genetic interactions to disease phenotypes has been particularly clear in cancer research, where an extreme genetic interaction, synthetic lethality, has been exploited as a therapeutic strategy. The obvious benefits of unmasking genetic background-specific vulnerabilities, coupled with the power of systematic genome editing, have fueled efforts to translate genetic interaction mapping from model organisms to human cells. Here, we review recent developments in genetic interaction mapping, with a focus on CRISPR-based genome editing technologies and cancer.

Integrating genetic and protein-protein interaction networks maps a functional wiring diagram of a cell

Systematic experimental approaches have led to construction of comprehensive genetic and protein-protein interaction networks for the budding yeast, Saccharomyces cerevisiae. Genetic interactions capture functional relationships between genes using phenotypic readouts, while protein-protein interactions identify physical connections between gene products. These complementary, and largely non-overlapping, networks provide a global view of the functional architecture of a cell, revealing general organizing principles, many of which appear to be evolutionarily conserved. Here, we focus on insights derived from the integration of large-scale genetic and protein-protein interaction networks, highlighting principles that apply to both unicellular and more complex systems, including human cells. Network integration reveals fundamental connections involving key functional modules of eukaryotic cells, defining a core network of cellular function, which could be elaborated to explore cell-type specificity in metazoans.

Systematic analysis of complex genetic interactions

To systematically explore complex genetic interactions, we constructed ~200,000 yeast triple mutants and scored negative trigenic interactions. We selected double-mutant query genes across a broad spectrum of biological processes, spanning a range of quantitative features of the global digenic interaction network and tested for a genetic interaction with a third mutation. Trigenic interactions often occurred among functionally related genes, and essential genes were hubs on the trigenic network. Despite their functional enrichment, trigenic interactions tended to link genes in distant bioprocesses and displayed a weaker magnitude than digenic interactions. We estimate that the global trigenic interaction network is ~100 times as large as the global digenic network, highlighting the potential for complex genetic interactions to affect the biology of inheritance, including the genotype-to-phenotype relationship.

Integrating images from multiple microscopy screens reveals diverse patterns of change in the subcellular localization of proteins

The evaluation of protein localization changes on a systematic level is a powerful tool for understanding how cells respond to environmental, chemical, or genetic perturbations. To date, work in understanding these proteomic responses through high-throughput imaging has catalogued localization changes independently for each perturbation. To distinguish changes that are targeted responses to the specific perturbation or more generalized programs, we developed a scalable approach to visualize the localization behavior of proteins across multiple experiments as a quantitative pattern. By applying this approach to 24 experimental screens consisting of nearly 400,000 images, we differentiated specific responses from more generalized ones, discovered nuance in the localization behavior of stress-responsive proteins, and formed hypotheses by clustering proteins that have similar patterns. Previous approaches aim to capture all localization changes for a single screen as accurately as possible, whereas our work aims to integrate large amounts of imaging data to find unexpected new cell biology.

The Candida albicans transcription factor Cas5 couples stress responses, drug resistance and cell cycle regulation

The capacity to coordinate environmental sensing with initiation of cellular responses underpins microbial survival and is crucial for virulence and stress responses in microbial pathogens. Here we define circuitry that enables the fungal pathogen Candida albicans to couple cell cycle dynamics with responses to cell wall stress induced by echinocandins, a front-line class of antifungal drugs. We discover that the C. albicans transcription factor Cas5 is crucial for proper cell cycle dynamics and responses to echinocandins, which inhibit β-1,3-glucan synthesis. Cas5 has distinct transcriptional targets under basal and stress conditions, is activated by the phosphatase Glc7, and can regulate the expression of target genes in concert with the transcriptional regulators Swi4 and Swi6. Thus, we illuminate a mechanism of transcriptional control that couples cell wall integrity with cell cycle regulation, and uncover circuitry governing antifungal drug resistance.Cas5 is a transcriptional regulator of responses to cell wall stress in the fungal pathogen Candida albicans. Here, Xie et al. show that Cas5 also modulates cell cycle dynamics and responses to antifungal drugs.

Mapping a diversity of genetic interactions in yeast

Genetic interactions occur when the combination of multiple mutations yields an unexpected phenotype, and they may confound our ability to fully understand the genetic mechanisms underlying complex diseases. Genetic interactions are challenging to study because there are millions of possible different variant combinations within a given genome. Consequently, they have primarily been systematically explored in unicellular model organisms, such as yeast, with a focus on pairwise genetic interactions between loss-of-function alleles. However, there are many different types of genetic interactions, such as those occurring between gain-of-function or heterozygous mutations. Here, we review recent advances made in the systematic analysis of such diverse genetic interactions in yeast, and briefly discuss how similar studies could be undertaken in human cells.

Evaluation and Design of Genome-Wide CRISPR/SpCas9 Knockout Screens

The adaptation of CRISPR/SpCas9 technology to mammalian cell lines is transforming the study of human functional genomics. Pooled libraries of CRISPR guide RNAs (gRNAs) targeting human protein-coding genes and encoded in viral vectors have been used to systematically create gene knockouts in a variety of human cancer and immortalized cell lines, in an effort to identify whether these knockouts cause cellular fitness defects. Previous work has shown that CRISPR screens are more sensitive and specific than pooled-library shRNA screens in similar assays, but currently there exists significant variability across CRISPR library designs and experimental protocols. In this study, we reanalyze 17 genome-scale knockout screens in human cell lines from three research groups, using three different genome-scale gRNA libraries. Using the Bayesian Analysis of Gene Essentiality algorithm to identify essential genes, we refine and expand our previously defined set of human core essential genes from 360 to 684 genes. We use this expanded set of reference core essential genes, CEG2, plus empirical data from six CRISPR knockout screens to guide the design of a sequence-optimized gRNA library, the Toronto KnockOut version 3.0 (TKOv3) library. We then demonstrate the high effectiveness of the library relative to reference sets of essential and nonessential genes, as well as other screens using similar approaches. The optimized TKOv3 library, combined with the CEG2 reference set, provide an efficient, highly optimized platform for performing and assessing gene knockout screens in human cell lines.

Mechanisms of suppression: The wiring of genetic resilience

Recent analysis of genome sequences has identified individuals that are healthy despite carrying severe disease-associated mutations. A possible explanation is that these individuals carry a second genomic perturbation that can compensate for the detrimental effects of the disease allele, a phenomenon referred to as suppression. In model organisms, suppression interactions are generally divided into two classes: genomic suppressors which are secondary mutations in the genome that bypass a mutant phenotype, and dosage suppression interactions in which overexpression of a suppressor gene rescues a mutant phenotype. Here, we describe the general properties of genomic and dosage suppression, with an emphasis on the budding yeast. We propose that suppression interactions between genetic variants are likely relevant for determining the penetrance of human traits. Consequently, an understanding of suppression mechanisms may guide the discovery of protective variants in healthy individuals that carry disease alleles, which could direct the rational design of new therapeutics.

Identifying pathogenicity of human variants via paralog-based yeast complementation

To better understand the health implications of personal genomes, we now face a largely unmet challenge to identify functional variants within disease-associated genes. Functional variants can be identified by trans-species complementation, e.g., by failure to rescue a yeast strain bearing a mutation in an orthologous human gene. Although orthologous complementation assays are powerful predictors of pathogenic variation, they are available for only a few percent of human disease genes. Here we systematically examine the question of whether complementation assays based on paralogy relationships can expand the number of human disease genes with functional variant detection assays. We tested over 1,000 paralogous human-yeast gene pairs for complementation, yielding 34 complementation relationships, of which 33 (97%) were novel. We found that paralog-based assays identified disease variants with success on par with that of orthology-based assays. Combining all homology-based assay results, we found that complementation can often identify pathogenic variants outside the homologous sequence region, presumably because of global effects on protein folding or stability. Within our search space, paralogy-based complementation more than doubled the number of human disease genes with a yeast-based complementation assay for disease variation.

Functional complementation assays of human disease-associated gene variants can reveal many more human disease variants at high confidence than current computational approaches, even using highly-diverged model organisms. However, this has generally only been possible for a minority of human disease genes for which orthologous complementation is known in the relevant model organism, so that alternative assays are urgently needed. Here we show that complementation relationships can be found for many additional human disease genes by exploiting paralogous human-yeast gene relationships, and that disease variant identification using paralogy-based assays performs on par with orthology-based assays.

Exploring genetic suppression interactions on a global scale

Genetic suppression occurs when the phenotypic defects caused by a mutation in a particular gene are rescued by a mutation in a second gene. To explore the principles of genetic suppression, we examined both literature-curated and unbiased experimental data, involving systematic genetic mapping and whole-genome sequencing, to generate a large-scale suppression network among yeast genes. Most suppression pairs identified novel relationships among functionally related genes, providing new insights into the functional wiring diagram of the cell. In addition to suppressor mutations, we identified frequent secondary mutations,in a subset of genes, that likely cause a delay in the onset of stationary phase, which appears to promote their enrichment within a propagating population. These findings allow us to formulate and quantify general mechanisms of genetic suppression.

Automated analysis of high-content microscopy data with deep learning

Existing computational pipelines for quantitative analysis of high‐content microscopy data rely on traditional machine learning approaches that fail to accurately classify more than a single dataset without substantial tuning and training, requiring extensive analysis. Here, we demonstrate that the application of deep learning to biological image data can overcome the pitfalls associated with conventional machine learning classifiers. Using a deep convolutional neural network (DeepLoc) to analyze yeast cell images, we show improved performance over traditional approaches in the automated classification of protein subcellular localization. We also demonstrate the ability of DeepLoc to classify highly divergent image sets, including images of pheromone‐arrested cells with abnormal cellular morphology, as well as images generated in different genetic backgrounds and in different laboratories. We offer an open‐source implementation that enables updating DeepLoc on new microscopy datasets. This study highlights deep learning as an important tool for the expedited analysis of high‐content microscopy data.

Machine learning and computer vision approaches for phenotypic profiling

Grys et al. review computer vision and machine-learning methods that have been applied to phenotypic profiling of image-based data. Descriptions are provided for segmentation, feature extraction, selection, and dimensionality reduction, as well as clustering, outlier detection, and classification of data.

With recent advances in high-throughput, automated microscopy, there has been an increased demand for effective computational strategies to analyze large-scale, image-based data. To this end, computer vision approaches have been applied to cell segmentation and feature extraction, whereas machine-learning approaches have been developed to aid in phenotypic classification and clustering of data acquired from biological images. Here, we provide an overview of the commonly used computer vision and machine-learning methods for generating and categorizing phenotypic profiles, highlighting the general biological utility of each approach.

Exploring Quantitative Yeast Phenomics with Single-Cell Analysis of DNA Damage Foci

A significant challenge of functional genomics is to develop methods for genome-scale acquisition and analysis of cell biological data. Here, we present an integrated method that combines genome-wide genetic perturbation of Saccharomyces cerevisiae with high-content screening to facilitate the genetic description of sub-cellular structures and compartment morphology. As proof of principle, we used a Rad52-GFP marker to examine DNA damage foci in ∼20 million single cells from ∼5,000 different mutant backgrounds in the context of selected genetic or chemical perturbations. Phenotypes were classified using a machine learning-based automated image analysis pipeline. 345 mutants were identified that had elevated numbers of DNA damage foci, almost half of which were identified only in sensitized backgrounds. Subsequent analysis of Vid22, a protein implicated in the DNA damage response, revealed that it acts together with the Sgs1 helicase at sites of DNA damage and preferentially binds G-quadruplex regions of the genome. This approach is extensible to numerous other cell biological markers and experimental systems.

Quantitative human cell encyclopedia

Quantitative cataloging of molecular entities in human cells will enable a better understanding of human physiology.

High-Content Screening for Quantitative Cell Biology

High-content screening (HCS), which combines automated fluorescence microscopy with quantitative image analysis, allows the acquisition of unbiased multiparametric data at the single cell level. This approach has been used to address diverse biological questions and identify a plethora of quantitative phenotypes of varying complexity in numerous different model systems. Here, we describe some recent applications of HCS, ranging from the identification of genes required for specific biological processes to the characterization of genetic interactions. We review the steps involved in the design of useful biological assays and automated image analysis, and describe major challenges associated with each. Additionally, we highlight emerging technologies and future challenges, and discuss how the field of HCS might be enhanced in the future.

High-Throughput Microscopy-Based Screening in Saccharomyces cerevisiae

The budding yeastSaccharomyces cerevisiaehas served as the pioneer model organism for virtually all genome-scale methods, including genome sequencing, DNA microarrays, gene deletion collections, and a variety of proteomic platforms. Yeast has also provided a test-bed for the development of systematic fluorescence-based imaging screens to enable the analysis of protein localization and abundance in vivo. Especially important has been the integration of high-throughput microscopy with automated image-processing methods, which has allowed researchers to overcome issues associated with manual image analysis and acquire unbiased, quantitative data. Here we provide an introduction to automated imaging in budding yeast.

An extended set of yeast-based functional assays accurately identifies human disease mutations

We can now routinely identify coding variants within individual human genomes. A pressing challenge is to determine which variants disrupt the function of disease-associated genes. Both experimental and computational methods exist to predict pathogenicity of human genetic variation. However, a systematic performance comparison between them has been lacking. Therefore, we developed and exploited a panel of 26 yeast-based functional complementation assays to measure the impact of 179 variants (101 disease- and 78 non-disease-associated variants) from 22 human disease genes. Using the resulting reference standard, we show that experimental functional assays in a 1-billion-year diverged model organism can identify pathogenic alleles with significantly higher precision and specificity than current computational methods.

Heritability and genetic basis of protein level variation in an outbred population

The genetic basis of heritable traits has been studied for decades. Although recent mapping efforts have elucidated genetic determinants of transcript levels, mapping of protein abundance has lagged. Here, we analyze levels of 4084 GFP-tagged yeast proteins in the progeny of a cross between a laboratory and a wild strain using flow cytometry and high-content microscopy. The genotype of trans variants contributed little to protein level variation between individual cells but explained >50% of the variance in the population’s average protein abundance for half of the GFP fusions tested. To map trans-acting factors responsible, we performed flow sorting and bulk segregant analysis of 25 proteins, finding a median of five protein quantitative trait loci (pQTLs) per GFP fusion. Further, we find that cis-acting variants predominate; the genotype of a gene and its surrounding region had a large effect on protein level six times more frequently than the rest of the genome combined. We present evidence for both shared and independent genetic control of transcript and protein abundance: More than half of the expression QTLs (eQTLs) contribute to changes in protein levels of regulated genes, but several pQTLs do not affect their cognate transcript levels. Allele replacements of genes known to underlie trans eQTL hotspots confirmed the correlation of effects on mRNA and protein levels. This study represents the first genome-scale measurement of genetic contribution to protein levels in single cells and populations, identifies more than a hundred trans pQTLs, and validates the propagation of effects associated with transcript variation to protein abundance.

eIF5A has a function in the cotranslational translocation of proteins into the ER

The putative eukaryotic translation initiation factor 5A (eIF5A) is a highly conserved and essential protein present in all organisms except bacteria. To be activated, eIF5A requires the conversion of a specific residue of lysine into hypusine. This hypusine modification occurs posttranslationally in two enzymatic steps, and the polyamine spermidine is the substrate. Despite having an essential function in translation elongation, the critical role played by eIF5A remains unclear. In addition to demonstrating genetic interactions with translation factors, eIF5A mutants genetically interact with mutations in YPT1, which encodes an essential protein involved in endoplasmic reticulum (ER)-to-Golgi vesicle transport. In this study, we investigated the correlation between the function of eIF5A in translation and secretion in yeast. The results of in vivo translocation assays and genetic interaction analyses suggest a specific role for eIF5A in the cotranslational translocation of proteins into the ER, but not in the posttranslational pathway. Additionally, we observed that a block in eIF5A activation up-regulates stress-induced chaperones, which also occurs when SRP function is lost. Finally, loss of eIF5A function affects binding of the ribosome-nascent chain complex to SRP. These results link eIF5A function in translation with a role of SRP in the cell and may help explain the dual effects of eIF5A in differential and general translation.

Unsupervised clustering of subcellular protein expression patterns in high-throughput microscopy images reveals protein complexes and functional relationships between proteins

Protein subcellular localization has been systematically characterized in budding yeast using fluorescently tagged proteins. Based on the fluorescence microscopy images, subcellular localization of many proteins can be classified automatically using supervised machine learning approaches that have been trained to recognize predefined image classes based on statistical features. Here, we present an unsupervised analysis of protein expression patterns in a set of high-resolution, high-throughput microscope images. Our analysis is based on 7 biologically interpretable features which are evaluated on automatically identified cells, and whose cell-stage dependency is captured by a continuous model for cell growth. We show that it is possible to identify most previously identified localization patterns in a cluster analysis based on these features and that similarities between the inferred expression patterns contain more information about protein function than can be explained by a previous manual categorization of subcellular localization. Furthermore, the inferred cell-stage associated to each fluorescence measurement allows us to visualize large groups of proteins entering the bud at specific stages of bud growth. These correspond to proteins localized to organelles, revealing that the organelles must be entering the bud in a stereotypical order. We also identify and organize a smaller group of proteins that show subtle differences in the way they move around the bud during growth. Our results suggest that biologically interpretable features based on explicit models of cell morphology will yield unprecedented power for pattern discovery in high-resolution, high-throughput microscopy images.

The location of a particular protein in the cell is one of the most important pieces of information that cell biologists use to understand its function. Fluorescent tags are a powerful way to determine the location of a protein in living cells. Nearly a decade ago, a collection of yeast strains was introduced, where in each strain a single protein was tagged with green fluorescent protein (GFP). Here, we show that by training a computer to accurately identify the buds of growing yeast cells, and then making simple fluorescence measurements in context of cell shape and cell stage, the computer could automatically discover most of the localization patterns (nucleus, cytoplasm, mitochondria, etc.) without any prior knowledge of what the patterns might be. Because we made the same, simple measurements for each yeast cell, we could compare and visualize the patterns of fluorescence for the entire collection of strains. This allowed us to identify large groups of proteins moving around the cell in a coordinated fashion, and to identify new, complex patterns that had previously been difficult to describe.

Global analysis of SUMO chain function reveals multiple roles in chromatin regulation

Multiple large-scale analyses in yeast implicate SUMO chain function in the maintenance of higher-order chromatin structure and transcriptional repression of environmental stress response genes.

Like ubiquitin, the small ubiquitin-related modifier (SUMO) proteins can form oligomeric “chains,” but the biological functions of these superstructures are not well understood. Here, we created mutant yeast strains unable to synthesize SUMO chains (smt3^allR) and subjected them to high-content microscopic screening, synthetic genetic array (SGA) analysis, and high-density transcript profiling to perform the first global analysis of SUMO chain function. This comprehensive assessment identified 144 proteins with altered localization or intensity in smt3^allR cells, 149 synthetic genetic interactions, and 225 mRNA transcripts (primarily consisting of stress- and nutrient-response genes) that displayed a >1.5-fold increase in expression levels. This information-rich resource strongly implicates SUMO chains in the regulation of chromatin. Indeed, using several different approaches, we demonstrate that SUMO chains are required for the maintenance of normal higher-order chromatin structure and transcriptional repression of environmental stress response genes in budding yeast.

SGAtools: one-stop analysis and visualization of array-based genetic interaction screens

Screening genome-wide sets of mutants for fitness defects provides a simple but powerful approach for exploring gene function, mapping genetic networks and probing mechanisms of drug action. For yeast and other microorganisms with global mutant collections, genetic or chemical-genetic interactions can be effectively quantified by growing an ordered array of strains on agar plates as individual colonies, and then scoring the colony size changes in response to a genetic or environmental perturbation. To do so, requires efficient tools for the extraction and analysis of quantitative data. Here, we describe SGAtools (http://sgatools.ccbr.utoronto.ca), a web-based analysis system for designer genetic screens. SGAtools outlines a series of guided steps that allow the user to quantify colony sizes from images of agar plates, correct for systematic biases in the observations and calculate a fitness score relative to a control experiment. The data can also be visualized online to explore the colony sizes on individual plates, view the distribution of resulting scores, highlight genes with the strongest signal and perform Gene Ontology enrichment analysis.

Bck2 acts through the MADS box protein Mcm1 to activate cell-cycle-regulated genes in budding yeast

The Bck2 protein is a potent genetic regulator of cell-cycle-dependent gene expression in budding yeast. To date, most experiments have focused on assessing a potential role for Bck2 in activation of the G1/S-specific transcription factors SBF (Swi4, Swi6) and MBF (Mbp1, Swi6), yet the mechanism of gene activation by Bck2 has remained obscure. We performed a yeast two-hybrid screen using a truncated version of Bck2 and discovered six novel Bck2-binding partners including Mcm1, an essential protein that binds to and activates M/G1 promoters through Early Cell cycle Box (ECB) elements as well as to G2/M promoters. At M/G1 promoters Mcm1 is inhibited by association with two repressors, Yox1 or Yhp1, and gene activation ensues once repression is relieved by an unknown activating signal. Here, we show that Bck2 interacts physically with Mcm1 to activate genes during G1 phase. We used chromatin immunoprecipitation (ChIP) experiments to show that Bck2 localizes to the promoters of M/G1-specific genes, in a manner dependent on functional ECB elements, as well as to the promoters of G1/S and G2/M genes. The Bck2-Mcm1 interaction requires valine 69 on Mcm1, a residue known to be required for interaction with Yox1. Overexpression of BCK2 decreases Yox1 localization to the early G1-specific CLN3 promoter and rescues the lethality caused by overexpression of YOX1. Our data suggest that Yox1 and Bck2 may compete for access to the Mcm1-ECB scaffold to ensure appropriate activation of the initial suite of genes required for cell cycle commitment.

Conserved rules govern genetic interaction degree across species

Background

Synthetic genetic interactions have recently been mapped on a genome scale in the budding yeast Saccharomyces cerevisiae, providing a functional view of the central processes of eukaryotic life. Currently, comprehensive genetic interaction networks have not been determined for other species, and we therefore sought to model conserved aspects of genetic interaction networks in order to enable the transfer of knowledge between species.

Results

Using a combination of physiological and evolutionary properties of genes, we built models that successfully predicted the genetic interaction degree of S. cerevisiae genes. Importantly, a model trained on S. cerevisiae gene features and degree also accurately predicted interaction degree in the fission yeast Schizosaccharomyces pombe, suggesting that many of the predictive relationships discovered in S. cerevisiae also hold in this evolutionarily distant yeast. In both species, high single mutant fitness defect, protein disorder, pleiotropy, protein-protein interaction network degree, and low expression variation were significantly predictive of genetic interaction degree. A comparison of the predicted genetic interaction degrees of S. pombe genes to the degrees of S. cerevisiae orthologs revealed functional rewiring of specific biological processes that distinguish these two species. Finally, predicted differences in genetic interaction degree were independently supported by differences in co-expression relationships of the two species.

Conclusions

Our findings show that there are common relationships between gene properties and genetic interaction network topology in two evolutionarily distant species. This conservation allows use of the extensively mapped S. cerevisiae genetic interaction network as an orthology-independent reference to guide the study of more complex species.

The budding yeast nuclear envelope adjacent to the nucleolus serves as a membrane sink during mitotic delay

The mechanisms that dictate nuclear shape are largely unknown. Here we screened the budding yeast deletion collection for mutants with abnormal nuclear shape. A common phenotype was the appearance of a nuclear extension, particularly in mutants in DNA repair and chromosome segregation genes. Our data suggest that these mutations led to the abnormal nuclear morphology indirectly, by causing a checkpoint-induced cell-cycle delay. Indeed, delaying cells in mitosis by other means also led to the appearance of nuclear extensions, whereas inactivating the DNA damage checkpoint pathway in a DNA repair mutant reduced the fraction of cells with nuclear extensions. Formation of a nuclear extension was specific to a mitotic delay, because cells arrested in S or G2 had round nuclei. Moreover, the nuclear extension always coincided with the nucleolus, while the morphology of the DNA mass remained largely unchanged. Finally, we found that phospholipid synthesis continued unperturbed when cells delayed in mitosis, and inhibiting phospholipid synthesis abolished the formation of nuclear extensions. Our data suggest a mechanism that promotes nuclear envelope expansion during mitosis. When mitotic progression is delayed, cells sequester the added membrane to the nuclear envelope associated with the nucleolus, possibly to avoid disruption of intranuclear organization.

Proteome-wide discovery of evolutionary conserved sequences in disordered regions

At least 30% of human proteins are thought to contain intrinsically disordered regions, which lack stable structural conformation. Despite lacking enzymatic functions and having few protein domains, disordered regions are functionally important for protein regulation and contain short linear motifs (short peptide sequences involved in protein-protein interactions), but in most disordered regions, the functional amino acid residues remain unknown. We searched for evolutionarily conserved sequences within disordered regions according to the hypothesis that conservation would indicate functional residues. Using a phylogenetic hidden Markov model (phylo-HMM), we made accurate, specific predictions of functional elements in disordered regions even when these elements are only two or three amino acids long. Among the conserved sequences that we identified were previously known and newly identified short linear motifs, and we experimentally verified key examples, including a motif that may mediate interaction between protein kinase Cbk1 and its substrates. We also observed that hub proteins, which interact with many partners in a protein interaction network, are highly enriched in these conserved sequences. Our analysis enabled the systematic identification of the functional residues in disordered regions and suggested that at least 5% of amino acids in disordered regions are important for function.

Restriction of histone gene transcription to S phase by phosphorylation of a chromatin boundary protein

The cell cycle-regulated expression of core histone genes is required for DNA replication and proper cell cycle progression in eukaryotic cells. Although some factors involved in histone gene transcription are known, the molecular mechanisms that ensure proper induction of histone gene expression during S phase remain enigmatic. Here we demonstrate that S-phase transcription of the model histone gene HTA1 in yeast is regulated by a novel attach-release mechanism involving phosphorylation of the conserved chromatin boundary protein Yta7 by both cyclin-dependent kinase 1 (Cdk1) and casein kinase 2 (CK2). Outside S phase, integrity of the AAA-ATPase domain is required for Yta7 boundary function, as defined by correct positioning of the histone chaperone Rtt106 and the chromatin remodeling complex RSC. Conversely, in S phase, Yta7 is hyperphosphorylated, causing its release from HTA1 chromatin and productive transcription. Most importantly, abrogation of Yta7 phosphorylation results in constitutive attachment of Yta7 to HTA1 chromatin, preventing efficient transcription post-recruitment of RNA polymerase II (RNAPII). Our study identified the chromatin boundary protein Yta7 as a key regulator that links S-phase kinases with RNAPII function at cell cycle-regulated histone gene promoters.

Functional wiring of the yeast kinome revealed by global analysis of genetic network motifs

A combinatorial genetic perturbation strategy was applied to interrogate the yeast kinome on a genome-wide scale. We assessed the global effects of gene overexpression or gene deletion to map an integrated genetic interaction network of synthetic dosage lethal (SDL) and loss-of-function genetic interactions (GIs) for 92 kinases, producing a meta-network of 8700 GIs enriched for pathways known to be regulated by cognate kinases. Kinases most sensitive to dosage perturbations had constitutive cell cycle or cell polarity functions under standard growth conditions. Condition-specific screens confirmed that the spectrum of kinase dosage interactions can be expanded substantially in activating conditions. An integrated network composed of systematic SDL, negative and positive loss-of-function GIs, and literature-curated kinase-substrate interactions revealed kinase-dependent regulatory motifs predictive of novel gene-specific phenotypes. Our study provides a valuable resource to unravel novel functional relationships and pathways regulated by kinases and outlines a general strategy for deciphering mutant phenotypes from large-scale GI networks.

Global functional map of the p23 molecular chaperone reveals an extensive cellular network

In parallel with evolutionary developments, the Hsp90 molecular chaperone system shifted from a simple prokaryotic factor into an expansive network that includes a variety of cochaperones. We have taken high-throughput genomic and proteomic approaches to better understand the abundant yeast p23 cochaperone Sba1. Our work revealed an unexpected p23 network that displayed considerable independence from known Hsp90 clients. Additionally, our data uncovered a broad nuclear role for p23, contrasting with the historical dogma of restricted cytosolic activities for molecular chaperones. Validation studies demonstrated that yeast p23 was required for proper Golgi function and ribosome biogenesis, and was necessary for efficient DNA repair from a wide range of mutagens. Notably, mammalian p23 had conserved roles in these pathways as well as being necessary for proper cell mobility. Taken together, our work demonstrates that the p23 chaperone serves a broad physiological network and functions both in conjunction with and sovereign to Hsp90.

Systematic exploration of synergistic drug pairs

Drug synergy allows a therapeutic effect to be achieved with lower doses of component drugs. Drug synergy can result when drugs target the products of genes that act in parallel pathways ('specific synergy'). Such cases of drug synergy should tend to correspond to synergistic genetic interaction between the corresponding target genes. Alternatively, 'promiscuous synergy' can arise when one drug non-specifically increases the effects of many other drugs, for example, by increased bioavailability. To assess the relative abundance of these drug synergy types, we examined 200 pairs of antifungal drugs in S. cerevisiae. We found 38 antifungal synergies, 37 of which were novel. While 14 cases of drug synergy corresponded to genetic interaction, 92% of the synergies we discovered involved only six frequently synergistic drugs. Although promiscuity of four drugs can be explained under the bioavailability model, the promiscuity of Tacrolimus and Pentamidine was completely unexpected. While many drug synergies correspond to genetic interactions, the majority of drug synergies appear to result from non-specific promiscuous synergy.

Exploring the conservation of synthetic lethal genetic interaction networks

High-throughput studies have enabled the large-scale mapping of synthetic lethal genetic interaction networks in the budding yeast Saccharomyces cerevisiae (S. cerevisiae). Recently, complementary high-throughput methods have been developed to map genetic interactions in the fission yeast Schizosaccharomyces pombe (S. pombe), enabling comparative analyses of genetic interaction networks between S. pombe and S. cerevisiae, two species separated by hundreds of millions of years of evolution. The resultant data has providing our first view of a possible core genetic interaction network shared between two distantly related eukaryotes, and identified numerous species-specific interactions that may contribute to the unique biology of these two different organisms. These and other results suggest that comparative interactomic studies will provide novel insights into the structure of genetic interaction networks.