A DNA copy number alteration classifier as a prognostic tool for prostate cancer patients.

BACKGROUND: Distinguishing between true indolent and potentially life-threatening prostate cancer is challenging in tumours displaying clinicopathologic features associated with low or intermediate risk of relapse. Several somatic DNA copy number alterations (CNAs) have been identified as potential prognostic biomarkers, but the standard cytogenetic method to assess them has a limited multiplexing capability. METHODS: Multiplex ligation-dependent probe amplification (MLPA) targeting 14 genes was optimised to survey 448 tumours of patients with low or intermediate risk (Grade Group 1-3, Gleason score ≤7) who underwent radical prostatectomy. A 6-gene CNA classifier was developed using random survival forest and Cox proportional hazard modelling to predict biochemical recurrence. RESULTS: The classifier score was significantly associated with biochemical recurrence after adjusting for standard clinicopathologic variables and the known prognostic index CAPRA-S score with a hazard ratio of 2.17 and 1.80, respectively (n = 406, P < 0.01). The prognostic value of this classifier was externally validated in published CNA data from three radical prostatectomy cohorts and one radiation therapy pre-treatment biopsy cohort. CONCLUSION: The 6-gene CNA classifier generated by a single MLPA assay compatible with the small quantities of DNA extracted from formalin-fixed paraffin-embedded (FFPE) tissue specimens has the potential to improve the clinical management of patients with low or intermediate risk disease.

novoBreak: local assembly for breakpoint detection in cancer genomes.

We present novoBreak, a genome-wide local assembly algorithm that discovers somatic and germline structural variation breakpoints in whole-genome sequencing data. novoBreak consistently outperformed existing algorithms on real cancer genome data and on synthetic tumors in the ICGC-TCGA DREAM 8.5 Somatic Mutation Calling Challenge primarily because it more effectively utilized reads spanning breakpoints. novoBreak also demonstrated great sensitivity in identifying short insertions and deletions.

Structural and Functional Characterization of a Caenorhabditis elegans Genetic Interaction Network within Pathways.

A genetic interaction (GI) is defined when the mutation of one gene modifies the phenotypic expression associated with the mutation of a second gene. Genome-wide efforts to map GIs in yeast revealed structural and functional properties of a GI network. This provided insights into the mechanisms underlying the robustness of yeast to genetic and environmental insults, and also into the link existing between genotype and phenotype. While a significant conservation of GIs and GI network structure has been reported between distant yeast species, such a conservation is not clear between unicellular and multicellular organisms. Structural and functional characterization of a GI network in these latter organisms is consequently of high interest. In this study, we present an in-depth characterization of ~1.5K GIs in the nematode Caenorhabditis elegans. We identify and characterize six distinct classes of GIs by examining a wide-range of structural and functional properties of genes and network, including co-expression, phenotypical manifestations, relationship with protein-protein interaction dense subnetworks (PDS) and pathways, molecular and biological functions, gene essentiality and pleiotropy. Our study shows that GI classes link genes within pathways and display distinctive properties, specifically towards PDS. It suggests a model in which pathways are composed of PDS-centric and PDS-independent GIs coordinating molecular machines through two specific classes of GIs involving pleiotropic and non-pleiotropic connectors. Our study provides the first in-depth characterization of a GI network within pathways of a multicellular organism. It also suggests a model to understand better how GIs control system robustness and evolution.

Chemical genomic screening of a Saccharomyces cerevisiae genomewide mutant collection reveals genes required for defense against four antimicrobial peptides derived from proteins found in human saliva.

To compare the effects of four antimicrobial peptides (MUC7 12-mer, histatin 12-mer, cathelicidin KR20, and a peptide containing lactoferricin amino acids 1 to 11) on the yeast Saccharomyces cerevisiae, we employed a genomewide fitness screen of combined collections of mutants with homozygous deletions of nonessential genes and heterozygous deletions of essential genes. When an arbitrary fitness score cutoffs of 1 (indicating a fitness defect, or hypersensitivity) and -1 (indicating a fitness gain, or resistance) was used, 425 of the 5,902 mutants tested exhibited altered fitness when treated with at least one peptide. Functional analysis of the 425 strains revealed enrichment among the identified deletions in gene groups associated with the Gene Ontology (GO) terms "ribosomal subunit," "ribosome biogenesis," "protein glycosylation," "vacuolar transport," "Golgi vesicle transport," "negative regulation of transcription," and others. Fitness profiles of all four tested peptides were highly similar, particularly among mutant strains exhibiting the greatest fitness defects. The latter group included deletions in several genes involved in induction of the RIM101 signaling pathway, including several components of the ESCRT sorting machinery. The RIM101 signaling regulates response of yeasts to alkaline and neutral pH and high salts, and our data indicate that this pathway also plays a prominent role in regulating protective measures against all four tested peptides. In summary, the results of the chemical genomic screens of S. cerevisiae mutant collection suggest that the four antimicrobial peptides, despite their differences in structure and physical properties, share many interactions with S. cerevisiae cells and consequently a high degree of similarity between their modes of action.

Identification of yeast genes that confer resistance to chitosan oligosaccharide (COS) using chemogenomics.

BACKGROUND: Chitosan oligosaccharide (COS), a deacetylated derivative of chitin, is an abundant, and renewable natural polymer. COS has higher antimicrobial properties than chitosan and is presumed to act by disrupting/permeabilizing the cell membranes of bacteria, yeast and fungi. COS is relatively non-toxic to mammals. By identifying the molecular and genetic targets of COS, we hope to gain a better understanding of the antifungal mode of action of COS. RESULTS: Three different chemogenomic fitness assays, haploinsufficiency (HIP), homozygous deletion (HOP), and multicopy suppression (MSP) profiling were combined with a transcriptomic analysis to gain insight in to the mode of action and mechanisms of resistance to chitosan oligosaccharides. The fitness assays identified 39 yeast deletion strains sensitive to COS and 21 suppressors of COS sensitivity. The genes identified are involved in processes such as RNA biology (transcription, translation and regulatory mechanisms), membrane functions (e.g. signalling, transport and targeting), membrane structural components, cell division, and proteasome processes. The transcriptomes of control wild type and 5 suppressor strains overexpressing ARL1, BCK2, ERG24, MSG5, or RBA50, were analyzed in the presence and absence of COS. Some of the up-regulated transcripts in the suppressor overexpressing strains exposed to COS included genes involved in transcription, cell cycle, stress response and the Ras signal transduction pathway. Down-regulated transcripts included those encoding protein folding components and respiratory chain proteins. The COS-induced transcriptional response is distinct from previously described environmental stress responses (i.e. thermal, salt, osmotic and oxidative stress) and pre-treatment with these well characterized environmental stressors provided little or any resistance to COS. CONCLUSIONS: Overexpression of the ARL1 gene, a member of the Ras superfamily that regulates membrane trafficking, provides protection against COS-induced cell membrane permeability and damage. We found that the ARL1 COS-resistant over-expression strain was as sensitive to Amphotericin B, Fluconazole and Terbinafine as the wild type cells and that when COS and Fluconazole are used in combination they act in a synergistic fashion. The gene targets of COS identified in this study indicate that COS's mechanism of action is different from other commonly studied fungicides that target membranes, suggesting that COS may be an effective fungicide for drug-resistant fungal pathogens.

Reverse genetics in Candida albicans predicts ARF cycling is essential for drug resistance and virulence.

Candida albicans, the major fungal pathogen of humans, causes life-threatening infections in immunocompromised individuals. Due to limited available therapy options, this can frequently lead to therapy failure and emergence of drug resistance. To improve current treatment strategies, we have combined comprehensive chemical-genomic screening in Saccharomyces cerevisiae and validation in C. albicans with the goal of identifying compounds that can couple with the fungistatic drug fluconazole to make it fungicidal. Among the genes identified in the yeast screen, we found that only AGE3, which codes for an ADP-ribosylation factor GTPase activating effector protein, abrogates fluconazole tolerance in C. albicans. The age3 mutant was more sensitive to other sterols and cell wall inhibitors, including caspofungin. The deletion of AGE3 in drug resistant clinical isolates and in constitutively active calcineurin signaling mutants restored fluconazole sensitivity. We confirmed chemically the AGE3-dependent drug sensitivity by showing a potent fungicidal synergy between fluconazole and brefeldin A (an inhibitor of the guanine nucleotide exchange factor for ADP ribosylation factors) in wild type C. albicans as well as in drug resistant clinical isolates. Addition of calcineurin inhibitors to the fluconazole/brefeldin A combination only initially improved pathogen killing. Brefeldin A synergized with different drugs in non-albicans Candida species as well as Aspergillus fumigatus. Microarray studies showed that core transcriptional responses to two different drug classes are not significantly altered in age3 mutants. The therapeutic potential of inhibiting ARF activities was demonstrated by in vivo studies that showed age3 mutants are avirulent in wild type mice, attenuated in virulence in immunocompromised mice and that fluconazole treatment was significantly more efficacious when ARF signaling was genetically compromised. This work describes a new, widely conserved, broad-spectrum mechanism involved in fungal drug resistance and virulence and offers a potential route for single or improved combination therapies.

Systematic Prediction of Antifungal Drug Synergy by Chemogenomic Screening in Saccharomyces cerevisiae.

Since the earliest days of using natural remedies, combining therapies for disease treatment has been standard practice. Combination treatments exhibit synergistic effects, broadly defined as a greater-than-additive effect of two or more therapeutic agents. Clinicians often use their experience and expertise to tailor such combinations to maximize the therapeutic effect. Although understanding and predicting biophysical underpinnings of synergy have benefitted from high-throughput screening and computational studies, one challenge is how to best design and analyze the results of synergy studies, especially because the number of possible combinations to test quickly becomes unmanageable. Nevertheless, the benefits of such studies are clear-by combining multiple drugs in the treatment of infectious disease and cancer, for instance, one can lessen host toxicity and simultaneously reduce the likelihood of resistance to treatment. This study introduces a new approach to characterize drug synergy, in which we extend the widely validated chemogenomic HIP-HOP assay to drug combinations; this assay involves parallel screening of comprehensive collections of barcoded deletion mutants. We identify a class of "combination-specific sensitive strains" that introduces mechanisms for the synergies we observe and further suggest focused follow-up studies.

Chemogenomic profiling predicts antifungal synergies.

Chemotherapies, HIV infections, and treatments to block organ transplant rejection are creating a population of immunocompromised individuals at serious risk of systemic fungal infections. Since single-agent therapies are susceptible to failure due to either inherent or acquired resistance, alternative therapeutic approaches such as multi-agent therapies are needed. We have developed a bioinformatics-driven approach that efficiently predicts compound synergy for such combinatorial therapies. The approach uses chemogenomic profiles in order to identify compound profiles that have a statistically significant degree of similarity to a fluconazole profile. The compounds identified were then experimentally verified to be synergistic with fluconazole and with each other, in both Saccharomyces cerevisiae and the fungal pathogen Candida albicans. Our method is therefore capable of accurately predicting compound synergy to aid the development of combinatorial antifungal therapies.

Comparison of small n statistical tests of differential expression applied to microarrays.

BACKGROUND: DNA microarrays provide data for genome wide patterns of expression between observation classes. Microarray studies often have small samples sizes, however, due to cost constraints or specimen availability. This can lead to poor random error estimates and inaccurate statistical tests of differential expression. We compare the performance of the standard t-test, fold change, and four small n statistical test methods designed to circumvent these problems. We report results of various normalization methods for empirical microarray data and of various random error models for simulated data. RESULTS: Three Empirical Bayes methods (CyberT, BRB, and limma t-statistics) were the most effective statistical tests across simulated and both 2-colour cDNA and Affymetrix experimental data. The CyberT regularized t-statistic in particular was able to maintain expected false positive rates with simulated data showing high variances at low gene intensities, although at the cost of low true positive rates. The Local Pooled Error (LPE) test introduced a bias that lowered false positive rates below theoretically expected values and had lower power relative to the top performers. The standard two-sample t-test and fold change were also found to be sub-optimal for detecting differentially expressed genes. The generalized log transformation was shown to be beneficial in improving results with certain data sets, in particular high variance cDNA data. CONCLUSION: Pre-processing of data influences performance and the proper combination of pre-processing and statistical testing is necessary for obtaining the best results. All three Empirical Bayes methods assessed in our study are good choices for statistical tests for small n microarray studies for both Affymetrix and cDNA data. Choice of method for a particular study will depend on software and normalization preferences.

Improved structural characterizations of the drkN SH3 domain unfolded state suggest a compact ensemble with native-like and non-native structure.

Due to their dynamic ensemble nature and a deficiency of experimental restraints, disordered states of proteins are difficult to characterize structurally. Here, we have expanded upon our previous work on the unfolded state of the Drosophila drk N-terminal (drkN) SH3 domain with our program ENSEMBLE, which assigns population weights to pregenerated conformers in order to calculate ensembles of structures whose properties are collectively consistent with experimental measurements. The experimental restraint set has been enlarged with newly measured paramagnetic relaxation enhancements from Cu(2+) bound to an amino terminal Cu(2+)-Ni(2+) binding (ATCUN) motif as well as nuclear Overhauser effect (NOE) and hydrogen exchange data from recent studies. In addition, two new pseudo-energy minimization algorithms have been implemented that have dramatically improved the speed of ENSEMBLE population weight assignment. Finally, we have greatly improved our conformational sampling by utilizing a variety of techniques to generate both random structures and structures that are biased to contain elements of native-like or non-native structure. Although it is not possible to uniquely define a representative structural ensemble, we have been able to assess various properties of the drkN SH3 domain unfolded state by performing ENSEMBLE minimizations of different conformer pools. Specifically, we have found that the experimental restraint set enforces a compact structural distribution that is not consistent with an overall native-like topology but shows preference for local non-native structure in the regions corresponding to the diverging turn and the beta5 strand of the folded state and for local native-like structure in the region corresponding to the beta6 and beta7 strands. We suggest that this approach could be generally useful for the structural characterization of disordered states.

Combining accurate tumor genome simulation with crowdsourcing to benchmark somatic structural variant detection.

BACKGROUND: The phenotypes of cancer cells are driven in part by somatic structural variants. Structural variants can initiate tumors, enhance their aggressiveness, and provide unique therapeutic opportunities. Whole-genome sequencing of tumors can allow exhaustive identification of the specific structural variants present in an individual cancer, facilitating both clinical diagnostics and the discovery of novel mutagenic mechanisms. A plethora of somatic structural variant detection algorithms have been created to enable these discoveries; however, there are no systematic benchmarks of them. Rigorous performance evaluation of somatic structural variant detection methods has been challenged by the lack of gold standards, extensive resource requirements, and difficulties arising from the need to share personal genomic information. RESULTS: To facilitate structural variant detection algorithm evaluations, we create a robust simulation framework for somatic structural variants by extending the BAMSurgeon algorithm. We then organize and enable a crowdsourced benchmarking within the ICGC-TCGA DREAM Somatic Mutation Calling Challenge (SMC-DNA). We report here the results of structural variant benchmarking on three different tumors, comprising 204 submissions from 15 teams. In addition to ranking methods, we identify characteristic error profiles of individual algorithms and general trends across them. Surprisingly, we find that ensembles of analysis pipelines do not always outperform the best individual method, indicating a need for new ways to aggregate somatic structural variant detection approaches. CONCLUSIONS: The synthetic tumors and somatic structural variant detection leaderboards remain available as a community benchmarking resource, and BAMSurgeon is available at https://github.com/adamewing/bamsurgeon .

Genes required for survival in microgravity revealed by genome-wide yeast deletion collections cultured during spaceflight.

Spaceflight is a unique environment with profound effects on biological systems including tissue redistribution and musculoskeletal stresses. However, the more subtle biological effects of spaceflight on cells and organisms are difficult to measure in a systematic, unbiased manner. Here we test the utility of the molecularly barcoded yeast deletion collection to provide a quantitative assessment of the effects of microgravity on a model organism. We developed robust hardware to screen, in parallel, the complete collection of ~4800 homozygous and ~5900 heterozygous (including ~1100 single-copy deletions of essential genes) yeast deletion strains, each carrying unique DNA that acts as strain identifiers. We compared strain fitness for the homozygous and heterozygous yeast deletion collections grown in spaceflight and ground, as well as plus and minus hyperosmolar sodium chloride, providing a second additive stressor. The genome-wide sensitivity profiles obtained from these treatments were then queried for their similarity to a compendium of drugs whose effects on the yeast collection have been previously reported. We found that the effects of spaceflight have high concordance with the effects of DNA-damaging agents and changes in redox state, suggesting mechanisms by which spaceflight may negatively affect cell fitness.

Mapping the cellular response to small molecules using chemogenomic fitness signatures.

Genome-wide characterization of the in vivo cellular response to perturbation is fundamental to understanding how cells survive stress. Identifying the proteins and pathways perturbed by small molecules affects biology and medicine by revealing the mechanisms of drug action. We used a yeast chemogenomics platform that quantifies the requirement for each gene for resistance to a compound in vivo to profile 3250 small molecules in a systematic and unbiased manner. We identified 317 compounds that specifically perturb the function of 121 genes and characterized the mechanism of specific compounds. Global analysis revealed that the cellular response to small molecules is limited and described by a network of 45 major chemogenomic signatures. Our results provide a resource for the discovery of functional interactions among genes, chemicals, and biological processes.

A high-throughput yeast assay identifies synergistic drug combinations.

Drug combinations are commonly used in the treatment of a range of diseases such as cancer, AIDS, and bacterial infections. Such combinations are less likely to be thwarted by resistance, and they have the desirable potential to be synergistic. Synergistic combinations can have decreased toxicity if lower doses of the constituent agents can be used. Conversely, antagonistic combinations can lead to lower efficacy of a treatment. Unfortunately, practical limitations, including the large number of possible combinations to be tested and the importance of optimizing concentrations and order of addition, discourage systematic studies of compound combinations. To address these limitations, we present a platform to screen drug combinations at multiple concentrations with varying orders of addition in Saccharomyces cerevisiae, at high throughput. In a proof of principle, we screened all possible pairwise combinations of 11 DNA damaging agents and found that of the 66 combinations tested, six were synergistic and three were antagonistic. The strength of two-thirds of these combinations was dependent on the order in which the drugs were added to the cells. We further tested the synergistic and antagonistic combinations in two cancer cell lines and found the combination of mitomycin C and irinotecan to be synergistic in both cell lines. This pilot study demonstrates the utility of using yeast for screening large matrices of drug combinations, and it provides a means to prioritize drug combination tests in human cells. Finally, we underscore the importance of testing the order of addition for assessing drug combinations.

Lysosomal disruption preferentially targets acute myeloid leukemia cells and progenitors.

Despite efforts to understand and treat acute myeloid leukemia (AML), there remains a need for more comprehensive therapies to prevent AML-associated relapses. To identify new therapeutic strategies for AML, we screened a library of on- and off-patent drugs and identified the antimalarial agent mefloquine as a compound that selectively kills AML cells and AML stem cells in a panel of leukemia cell lines and in mice. Using a yeast genome-wide functional screen for mefloquine sensitizers, we identified genes associated with the yeast vacuole, the homolog of the mammalian lysosome. Consistent with this, we determined that mefloquine disrupts lysosomes, directly permeabilizes the lysosome membrane, and releases cathepsins into the cytosol. Knockdown of the lysosomal membrane proteins LAMP1 and LAMP2 resulted in decreased cell viability, as did treatment of AML cells with known lysosome disrupters. Highlighting a potential therapeutic rationale for this strategy, leukemic cells had significantly larger lysosomes compared with normal cells, and leukemia-initiating cells overexpressed lysosomal biogenesis genes. These results demonstrate that lysosomal disruption preferentially targets AML cells and AML progenitor cells, providing a rationale for testing lysosomal disruption as a novel therapeutic strategy for AML.

Comparative chemogenomics to examine the mechanism of action of dna-targeted platinum-acridine anticancer agents.

Platinum-based drugs have been used to successfully treat diverse cancers for several decades. Cisplatin, the original compound of this class, cross-links DNA, resulting in cell cycle arrest and cell death via apoptosis. Cisplatin is effective against several tumor types, yet it exhibits toxic side effects and tumors often develop resistance. To mitigate these liabilities while maintaining potency, we generated a library of non-classical platinum-acridine hybrid agents and assessed their mechanisms of action using a validated genome-wide screening approach in Saccharomyces cerevisiae and in the distantly related yeast Schizosaccharomyces pombe. Chemogenomic profiles from both S. cerevisiae and S. pombe demonstrate that several of the platinum-acridines damage DNA differently than cisplatin based on their requirement for distinct modules of DNA repair.

Dissecting DNA damage response pathways by analysing protein localization and abundance changes during DNA replication stress.

Relocalization of proteins is a hallmark of the DNA damage response. We use high-throughput microscopic screening of the yeast GFP fusion collection to develop a systems-level view of protein reorganization following drug-induced DNA replication stress. Changes in protein localization and abundance reveal drug-specific patterns of functional enrichments. Classification of proteins by subcellular destination enables the identification of pathways that respond to replication stress. We analysed pairwise combinations of GFP fusions and gene deletion mutants to define and order two previously unknown DNA damage responses. In the first, Cmr1 forms subnuclear foci that are regulated by the histone deacetylase Hos2 and are distinct from the typical Rad52 repair foci. In a second example, we find that the checkpoint kinases Mec1/Tel1 and the translation regulator Asc1 regulate P-body formation. This method identifies response pathways that were not detected in genetic and protein interaction screens, and can be readily applied to any form of chemical or genetic stress to reveal cellular response pathways.

Functional analysis with a barcoder yeast gene overexpression system.

Systematic analysis of gene overexpression phenotypes provides an insight into gene function, enzyme targets, and biological pathways. Here, we describe a novel functional genomics platform that enables a highly parallel and systematic assessment of overexpression phenotypes in pooled cultures. First, we constructed a genome-level collection of ~5100 yeast barcoder strains, each of which carries a unique barcode, enabling pooled fitness assays with a barcode microarray or sequencing readout. Second, we constructed a yeast open reading frame (ORF) galactose-induced overexpression array by generating a genome-wide set of yeast transformants, each of which carries an individual plasmid-born and sequence-verified ORF derived from the Saccharomyces cerevisiae full-length EXpression-ready (FLEX) collection. We combined these collections genetically using synthetic genetic array methodology, generating ~5100 strains, each of which is barcoded and overexpresses a specific ORF, a set we termed "barFLEX." Additional synthetic genetic array allows the barFLEX collection to be moved into different genetic backgrounds. As a proof-of-principle, we describe the properties of the barFLEX overexpression collection and its application in synthetic dosage lethality studies under different environmental conditions.

Searching for signaling balance through the identification of genetic interactors of the Rab guanine-nucleotide dissociation inhibitor gdi-1.

BACKGROUND: The symptoms of numerous diseases result from genetic mutations that disrupt the homeostasis maintained by the appropriate integration of signaling gene activities. The relationships between signaling genes suggest avenues through which homeostasis can be restored and disease symptoms subsequently reduced. Specifically, disease symptoms caused by loss-of-function mutations in a particular gene may be reduced by concomitant perturbations in genes with antagonistic activities. METHODOLOGY/PRINCIPAL FINDINGS: Here we use network-neighborhood analyses to predict genetic interactions in Caenorhabditis elegans towards mapping antagonisms and synergisms between genes in an animal model. Most of the predicted interactions are novel, and the experimental validation establishes that our approach provides a gain in accuracy compared to previous efforts. In particular, we identified genetic interactors of gdi-1, the orthologue of GDI1, a gene associated with mental retardation in human. Interestingly, some gdi-1 interactors have human orthologues with known neurological functions, and upon validation of the interactions in mammalian systems, these orthologues would be potential therapeutic targets for GDI1-associated neurological disorders. We also observed the conservation of a gdi-1 interaction between different cellular systems in C. elegans, suggesting the involvement of GDI1 in human muscle degeneration. CONCLUSIONS/SIGNIFICANCE: We developed a novel predictor of genetic interactions that may have the ability to significantly streamline the identification of therapeutic targets for monogenic disorders involving genes conserved between human and C. elegans.

A biochemical genomics screen for substrates of Ste20p kinase enables the in silico prediction of novel substrates.

The Ste20/PAK family is involved in many cellular processes, including the regulation of actin-based cytoskeletal dynamics and the activation of MAPK signaling pathways. Despite its numerous roles, few of its substrates have been identified. To better characterize the roles of the yeast Ste20p kinase, we developed an in vitro biochemical genomics screen to identify its substrates. When applied to 539 purified yeast proteins, the screen reported 14 targets of Ste20p phosphorylation. We used the data resulting from our screen to build an in silico predictor to identify Ste20p substrates on a proteome-wide basis. Since kinase-substrate specificity is often mediated by additional binding events at sites distal to the phosphorylation site, the predictor uses the presence/absence of multiple sequence motifs to evaluate potential substrates. Statistical validation estimates a threefold improvement in substrate recovery over random predictions, despite the lack of a single dominant motif that can characterize Ste20p phosphorylation. The set of predicted substrates significantly overrepresents elements of the genetic and physical interaction networks surrounding Ste20p, suggesting that some of the predicted substrates are in vivo targets. We validated this combined experimental and computational approach for identifying kinase substrates by confirming the in vitro phosphorylation of polarisome components Bni1p and Bud6p, thus suggesting a mechanism by which Ste20p effects polarized growth.