Sparse clusterability: testing for cluster structure in high dimensions.

BACKGROUND: Cluster analysis is utilized frequently in scientific theory and applications to separate data into groups. A key assumption in many clustering algorithms is that the data was generated from a population consisting of multiple distinct clusters. Clusterability testing allows users to question the inherent assumption of latent cluster structure, a theoretical requirement for meaningful results in cluster analysis. RESULTS: This paper proposes methods for clusterability testing designed for high-dimensional data by utilizing sparse principal component analysis. Type I error and power of the clusterability tests are evaluated using simulated data with different types of cluster structure in high dimensions. Empirical performance of the new methods is evaluated and compared with prior methods on gene expression, microarray, and shotgun proteomics data. Our methods had reasonably low Type I error and maintained power for many datasets with a variety of structures and dimensions. Cluster structure was not detectable in other datasets with spatially close clusters. CONCLUSION: This is the first analysis of clusterability testing on both simulated and real-world high-dimensional data.

Targeted Therapy Given after Anti-PD-1 Leads to Prolonged Responses in Mouse Melanoma Models through Sustained Antitumor Immunity.

Immunotherapy (IT) and targeted therapy (TT) are both effective against melanoma, but their combination is frequently toxic. Here, we investigated whether the sequence of IT (anti-PD-1)→ TT (ceritinib-trametinib or dabrafenib-trametinib) was associated with improved antitumor responses in mouse models of BRAF- and NRAS-mutant melanoma. Mice with NRAS-mutant (SW1) or BRAF-mutant (SM1) mouse melanomas were treated with either IT, TT, or the sequence of IT→TT. Tumor volumes were measured, and samples from the NRAS-mutant melanomas were collected for immune-cell analysis, single-cell RNA sequencing (scRNA-seq), and reverse phase protein analysis (RPPA). scRNA-seq demonstrated that the IT→TT sequence modulated the immune environment, leading to increased infiltration of T cells, monocytes, dendritic cells and natural killer cells, and decreased numbers of tumor-associated macrophages, myeloid-derived suppressor cells, and regulatory T cells. Durable responses to the IT→TT sequence were dependent on T-cell activity, with depletion of CD8+, but not CD4+, T cells abrogating the therapeutic response. An analysis of transcriptional heterogeneity in the melanoma compartment showed the sequence of IT→TT enriched for a population of melanoma cells with increased expression of MHC class I and melanoma antigens. RPPA analysis demonstrated that the sustained immune response induced by IT→TT suppressed tumor-intrinsic signaling pathways required for therapeutic escape. These studies establish that upfront IT improves the responses to TT in BRAF- and NRAS-mutant melanoma models.

Predictive values, uncertainty, and interpretation of serology tests for the novel coronavirus.

Antibodies testing in the coronavirus era is frequently promoted, but the underlying statistics behind their validation has come under more scrutiny in recent weeks. We provide calculations, interpretations, and plots of positive and negative predictive values under a variety of scenarios. Prevalence, sensitivity, and specificity are estimated within ranges of values from researchers and antibodies manufacturers. Illustrative examples are highlighted, and interactive plots are provided in the Supplementary Material. Implications are discussed for society overall and across diverse locations with different levels of disease burden. Specifically, the proportion of positive serology tests that are false can differ drastically from up to 3% to 88% for people from different places with different proportions of infected people in the populations while the false negative rate is typically under 10%.

HDAC8 Regulates a Stress Response Pathway in Melanoma to Mediate Escape from BRAF Inhibitor Therapy.

Melanoma cells have the ability to switch to a dedifferentiated, invasive phenotype in response to multiple stimuli. Here, we show that exposure of melanomas to multiple stresses including BRAF-MEK inhibitor therapy, hypoxia, and UV irradiation leads to an increase in histone deacetylase 8 (HDAC8) activity and the adoption of a drug-resistant phenotype. Mass spectrometry-based phosphoproteomics implicated HDAC8 in the regulation of MAPK and AP-1 signaling. Introduction of HDAC8 into drug-naïve melanoma cells conveyed resistance both in vitro and in vivo. HDAC8-mediated BRAF inhibitor resistance was mediated via receptor tyrosine kinase activation, leading to MAPK signaling. Although HDACs function at the histone level, they also regulate nonhistone substrates, and introduction of HDAC8 decreased the acetylation of c-Jun, increasing its transcriptional activity and enriching for an AP-1 gene signature. Mutation of the putative c-Jun acetylation site at lysine 273 increased transcriptional activation of c-Jun in melanoma cells and conveyed resistance to BRAF inhibition. In vivo xenograft studies confirmed the key role of HDAC8 in therapeutic adaptation, with both nonselective and HDAC8-specific inhibitors enhancing the durability of BRAF inhibitor therapy. Our studies demonstrate that HDAC8-specific inhibitors limit the adaptation of melanoma cells to multiple stresses including BRAF-MEK inhibition. SIGNIFICANCE: This study provides evidence that HDAC8 drives transcriptional plasticity in melanoma cells in response to a range of stresses through direct deacetylation of c-Jun.Graphical Abstract: http://cancerres.aacrjournals.org/content/canres/79/11/2947/F1.large.jpg.

A Transcriptome-Wide Association Study Among 97,898 Women to Identify Candidate Susceptibility Genes for Epithelial Ovarian Cancer Risk.

Large-scale genome-wide association studies (GWAS) have identified approximately 35 loci associated with epithelial ovarian cancer (EOC) risk. The majority of GWAS-identified disease susceptibility variants are located in noncoding regions, and causal genes underlying these associations remain largely unknown. Here, we performed a transcriptome-wide association study to search for novel genetic loci and plausible causal genes at known GWAS loci. We used RNA sequencing data (68 normal ovarian tissue samples from 68 individuals and 6,124 cross-tissue samples from 369 individuals) and high-density genotyping data from European descendants of the Genotype-Tissue Expression (GTEx V6) project to build ovarian and cross-tissue models of genetically regulated expression using elastic net methods. We evaluated 17,121 genes for their cis-predicted gene expression in relation to EOC risk using summary statistics data from GWAS of 97,898 women, including 29,396 EOC cases. With a Bonferroni-corrected significance level of P < 2.2 × 10-6, we identified 35 genes, including FZD4 at 11q14.2 (Z = 5.08, P = 3.83 × 10-7, the cross-tissue model; 1 Mb away from any GWAS-identified EOC risk variant), a potential novel locus for EOC risk. All other 34 significantly associated genes were located within 1 Mb of known GWAS-identified loci, including 23 genes at 6 loci not previously linked to EOC risk. Upon conditioning on nearby known EOC GWAS-identified variants, the associations for 31 genes disappeared and three genes remained (P < 1.47 × 10-3). These data identify one novel locus (FZD4) and 34 genes at 13 known EOC risk loci associated with EOC risk, providing new insights into EOC carcinogenesis.Significance: Transcriptomic analysis of a large cohort confirms earlier GWAS loci and reveals FZD4 as a novel locus associated with EOC risk. Cancer Res; 78(18); 5419-30. ©2018 AACR.

Dabrafenib inhibits the growth of BRAF-WT cancers through CDK16 and NEK9 inhibition.

Although the BRAF inhibitors dabrafenib and vemurafenib have both proven successful against BRAF-mutant melanoma, there seem to be differences in their mechanisms of action. Here, we show that dabrafenib is more effective at inhibiting the growth of NRAS-mutant and KRAS-mutant cancer cell lines than vemurafenib. Using mass spectrometry-based chemical proteomics, we identified NEK9 and CDK16 as unique targets of dabrafenib. Both NEK9 and CDK16 were highly expressed in specimens of advanced melanoma, with high expression of both proteins correlating with a worse overall survival. A role for NEK9 in the growth of NRAS- and KRAS-mutant cell lines was suggested by siRNA studies in which silencing was associated with decreased proliferation, cell cycle arrest associated with increased p21 expression, inhibition of phospho-CHK1, decreased CDK4 expression, and the initiation of a senescence response. Inhibition of CDK4 but not CHK1 recapitulated the effects of NEK9 silencing, indicating this to be the likely mechanism of growth inhibition. We next turned our attention to CDK16 and found that its knockdown inhibited the phosphorylation of the Rb protein at S780 and increased expression of p27. Both of these effects were phenocopied in NRAS- and KRAS-mutant cancer cells by dabrafenib, but not vemurafenib. Combined silencing of NEK9 and CDK16 was associated with enhanced inhibition of melanoma cell proliferation. In summary, we have identified dabrafenib as a potent inhibitor of NEK9 and CDK16, and our studies suggest that inhibition of these kinases may have activity against cancers that do not harbor BRAF mutations.

Improved survival of patients with melanoma brain metastases in the era of targeted BRAF and immune checkpoint therapies.

BACKGROUND: The development of brain metastases is common for systemic treatment failure in patients with melanoma and has been associated with a poor prognosis. Recent advances with BRAF and immune checkpoint therapies have led to improved patient survival. Herein, the authors evaluated the risk of de novo brain metastases and survival among patients with melanoma brain metastases (MBM) since the introduction of more effective therapies. METHODS: Patients with unresectable AJCC stage III/IV melanoma who received first-line systemic therapy at Moffitt Cancer Center between 2000 and 2012 were identified. Data were collected regarding patient characteristics, stage of disease, systemic therapies, MBM status/management, and overall survival (OS). The risk of de novo MBM was calculated using a generalized estimating equation model and survival comparisons were performed using Kaplan-Meier and Cox proportional analyses. RESULTS: A total of 610 patients were included, 243 of whom were diagnosed with MBM (40%). Patients with MBM were younger, with a lower frequency of regional metastasis. No significant differences were noted with regard to sex, BRAF status, or therapeutic class. The risk of de novo MBM was found to be similar among patients treated with chemotherapy, biochemotherapy, BRAF-targeted therapy, ipilimumab, and anti-programmed cell death protein 1/programmed death-ligand 1 regimens. The median OS of patients with MBM was significantly shorter when determined from the time of first regional/distant metastasis but not when determined from the time of first systemic therapy. The median OS from the time of MBM diagnosis was 7.5 months, 8.5 months, and 22.7 months, respectively, for patients diagnosed from 2000 to 2008, 2009 to 2010, and 2011 to the time of last follow-up (P = .002). CONCLUSIONS: Brain metastases remain a common source of systemic treatment failure. The OS for patients with MBM has improved significantly. Further research into MBM prevention is needed. Cancer 2018;124:297-305. © 2017 American Cancer Society.

Multiplexed Liquid Chromatography-Multiple Reaction Monitoring Mass Spectrometry Quantification of Cancer Signaling Proteins.

Quantitative evaluation of protein expression across multiple cancer-related signaling pathways (e.g., Wnt/ß-catenin, TGF-ß, receptor tyrosine kinases (RTK), MAP kinases, NF-κB, and apoptosis) in tumor tissues may enable the development of a molecular profile for each individual tumor that can aid in the selection of appropriate targeted cancer therapies. Here, we describe the development of a broadly applicable protocol to develop and implement quantitative mass spectrometry assays using cell line models and frozen tissue specimens from colon cancer patients. Cell lines are used to develop peptide-based assays for protein quantification, which are incorporated into a method based on SDS-PAGE protein fractionation, in-gel digestion, and liquid chromatography-multiple reaction monitoring mass spectrometry (LC-MRM/MS). This analytical platform is then applied to frozen tumor tissues. This protocol can be broadly applied to the study of human disease using multiplexed LC-MRM assays.

Target Identification in Small Cell Lung Cancer via Integrated Phenotypic Screening and Activity-Based Protein Profiling.

To overcome hurdles in identifying key kinases in small cell lung cancer (SCLC), we integrated a target-agnostic phenotypic screen of kinase inhibitors with target identification using activity-based protein profiling (ABPP) in which a desthiobiotin-ATP probe was used. We screened 21 SCLC cell lines with known c-MYC amplification status for alterations in viability using a chemical library of 235 small-molecule kinase inhibitors. One screen hit compound was interrogated with ABPP, and, through this approach, we reidentified Aurora kinase B as a critical kinase in MYC-amplified SCLC cells. We next extended the platform to a second compound that had activity in SCLC cell lines lacking c-MYC amplification and identified TANK-binding kinase 1, a kinase that affects cell viability, polo-like kinase-1 signaling, G2-M arrest, and apoptosis in SCLC cells lacking MYC amplification. These results demonstrate that phenotypic screening combined with ABPP can identify key disease drivers, suggesting that this approach, which combines new chemical probes and disease cell screens, has the potential to identify other important targets in other cancer types. Mol Cancer Ther; 15(2); 334-42. ©2016 AACR.

INCORPORATING BIOLOGICAL INFORMATION INTO LINEAR MODELS: A BAYESIAN APPROACH TO THE SELECTION OF PATHWAYS AND GENES.

The vast amount of biological knowledge accumulated over the years has allowed researchers to identify various biochemical interactions and define different families of pathways. There is an increased interest in identifying pathways and pathway elements involved in particular biological processes. Drug discovery efforts, for example, are focused on identifying biomarkers as well as pathways related to a disease. We propose a Bayesian model that addresses this question by incorporating information on pathways and gene networks in the analysis of DNA microarray data. Such information is used to define pathway summaries, specify prior distributions, and structure the MCMC moves to fit the model. We illustrate the method with an application to gene expression data with censored survival outcomes. In addition to identifying markers that would have been missed otherwise and improving prediction accuracy, the integration of existing biological knowledge into the analysis provides a better understanding of underlying molecular processes.

A BAYESIAN GRAPHICAL MODELING APPROACH TO MICRORNA REGULATORY NETWORK INFERENCE.

It has been estimated that about 30% of the genes in the human genome are regulated by microRNAs (miRNAs). These are short RNA sequences that can down-regulate the levels of mRNAs or proteins in animals and plants. Genes regulated by miRNAs are called targets. Typically, methods for target prediction are based solely on sequence data and on the structure information. In this paper we propose a Bayesian graphical modeling approach that infers the miRNA regulatory network by integrating expression levels of miRNAs with their potential mRNA targets and, via the prior probability model, with their sequence/structure information. We use a directed graphical model with a particular structure adapted to our data based on biological considerations. We then achieve network inference using stochastic search methods for variable selection that allow us to explore the huge model space via MCMC. A time-dependent coefficients model is also implemented. We consider experimental data from a study on a very well-known developmental toxicant causing neural tube defects, hyperthermia. Some of the pairs of target gene and miRNA we identify seem very plausible and warrant future investigation. Our proposed method is general and can be easily applied to other types of network inference by integrating multiple data sources.

Gene and microRNA expression in p53-deficient day 8.5 mouse embryos.

BACKGROUND: Neural tube defects (NTDs) are one of the most common human birth defects, with a prevalence of approximately 1 in 1000 live births in the United States. In animal studies, deletion of p53 leads to a significant increase in embryos that exhibit exencephaly. Whereas several studies have closely investigated the morphologic changes of p53-deficient embryos, no study has reported the molecular-level alteration in p53-deficient embryos. Here we attempt to identify genes and microRNAs (miRNAs) modified by deletion of p53 in day 8.5 mouse embryos. METHODS: Mouse embryos from p53 heterozygous crosses were collected, genotyped, and embryos of similar genotype (+/+; +/-; -/-) were pooled. RNA from the pooled samples was isolated to determine mRNA and miRNA expression levels using Whole Genome Bioarrays and Low Density Arrays, respectively. RESULTS: In p53 -/- embryos, 388 genes showed statistically significant alteration in gene expression of more than twofold compared to p53 +/+ embryos. Expression of p53 and well known p53 target genes, such as p21 and cyclin G1, were significantly down-regulated in p53 -/- embryos. In contrast, expression of other p53 target genes, such as Mdm2, Noxa, and Puma, were unchanged. We also identified six genes (Csk, Itga3, Jarid2, Prkaca, Rarg, and Sall4), known to cause NTDs when deleted, that are also down-regulated in p53 -/- embryos. Finally, five miRNAs (mir-1, mir-30e-3p, mir-142-3p, mir-301, and mir-331) also showed statistically significant alterations in expression levels in p53 -/- embryos compared to p53 +/+ embryos. Combined analysis of the experimental data using stepwise regression model and two publicly available algorithms identified putative target genes of these miRNAs. CONCLUSIONS: Our data have identified genes and miRNAs that may be involved in the mechanisms underlining NTDs and begin to define the developmental role of p53 in the etiology of NTDs.

A cDNA microarray for Crassostrea virginica and C. gigas.

The eastern oyster, Crassostrea virginica, and the Pacific oyster, C. gigas, are species of global economic significance as well as important components of estuarine ecosystems and models for genetic and environmental studies. To enhance the molecular tools available for oyster research, an international group of collaborators has constructed a 27,496-feature cDNA microarray containing 4460 sequences derived from C. virginica, 2320 from C. gigas, and 16 non-oyster DNAs serving as positive and negative controls. The performance of the array was assessed by gene expression profiling using gill and digestive gland RNA derived from both C. gigas and C. virginica, and digestive gland RNA from C. ariakensis. The utility of the microarray for detection of homologous genes by cross-hybridization between species was also assessed and the correlation between hybridization intensity and sequence homology for selected genes determined. The oyster cDNA microarray is publicly available to the research community on a cost-recovery basis.

A multivariate prediction model for microarray cross-hybridization.

BACKGROUND: Expression microarray analysis is one of the most popular molecular diagnostic techniques in the post-genomic era. However, this technique faces the fundamental problem of potential cross-hybridization. This is a pervasive problem for both oligonucleotide and cDNA microarrays; it is considered particularly problematic for the latter. No comprehensive multivariate predictive modeling has been performed to understand how multiple variables contribute to (cross-) hybridization. RESULTS: We propose a systematic search strategy using multiple multivariate models [multiple linear regressions, regression trees, and artificial neural network analyses (ANNs)] to select an effective set of predictors for hybridization. We validate this approach on a set of DNA microarrays with cytochrome p450 family genes. The performance of our multiple multivariate models is compared with that of a recently proposed third-order polynomial regression method that uses percent identity as the sole predictor. All multivariate models agree that the 'most contiguous base pairs between probe and target sequences,' rather than percent identity, is the best univariate predictor. The predictive power is improved by inclusion of additional nonlinear effects, in particular target GC content, when regression trees or ANNs are used. CONCLUSION: A systematic multivariate approach is provided to assess the importance of multiple sequence features for hybridization and of relationships among these features. This approach can easily be applied to larger datasets. This will allow future developments of generalized hybridization models that will be able to correct for false-positive cross-hybridization signals in expression experiments.

Marine genomics: a clearing-house for genomic and transcriptomic data of marine organisms.

BACKGROUND: The Marine Genomics project is a functional genomics initiative developed to provide a pipeline for the curation of Expressed Sequence Tags (ESTs) and gene expression microarray data for marine organisms. It provides a unique clearing-house for marine specific EST and microarray data and is currently available at http://www.marinegenomics.org. DESCRIPTION: The Marine Genomics pipeline automates the processing, maintenance, storage and analysis of EST and microarray data for an increasing number of marine species. It currently contains 19 species databases (over 46,000 EST sequences) that are maintained by registered users from local and remote locations in Europe and South America in addition to the USA. A collection of analysis tools are implemented. These include a pipeline upload tool for EST FASTA file, sequence trace file and microarray data, an annotative text search, automated sequence trimming, sequence quality control (QA/QC) editing, sequence BLAST capabilities and a tool for interactive submission to GenBank. Another feature of this resource is the integration with a scientific computing analysis environment implemented by MATLAB. CONCLUSION: The conglomeration of multiple marine organisms with integrated analysis tools enables users to focus on the comprehensive descriptions of transcriptomic responses to typical marine stresses. This cross species data comparison and integration enables users to contain their research within a marine-oriented data management and analysis environment.

Optimal cDNA microarray design using expressed sequence tags for organisms with limited genomic information.

BACKGROUND: Expression microarrays are increasingly used to characterize environmental responses and host-parasite interactions for many different organisms. Probe selection for cDNA microarrays using expressed sequence tags (ESTs) is challenging due to high sequence redundancy and potential cross-hybridization between paralogous genes. In organisms with limited genomic information, like marine organisms, this challenge is even greater due to annotation uncertainty. No general tool is available for cDNA microarray probe selection for these organisms. Therefore, the goal of the design procedure described here is to select a subset of ESTs that will minimize sequence redundancy and characterize potential cross-hybridization while providing functionally representative probes. RESULTS: Sequence similarity between ESTs, quantified by the E-value of pair-wise alignment, was used as a surrogate for expected hybridization between corresponding sequences. Using this value as a measure of dissimilarity, sequence redundancy reduction was performed by hierarchical cluster analyses. The choice of how many microarray probes to retain was made based on an index developed for this research: a sequence diversity index (SDI) within a sequence diversity plot (SDP). This index tracked the decreasing within-cluster sequence diversity as the number of clusters increased. For a given stage in the agglomeration procedure, the EST having the highest similarity to all the other sequences within each cluster, the centroid EST, was selected as a microarray probe. A small dataset of ESTs from Atlantic white shrimp (Litopenaeus setiferus) was used to test this algorithm so that the detailed results could be examined. The functional representative level of the selected probes was quantified using Gene Ontology (GO) annotations. CONCLUSIONS: For organisms with limited genomic information, combining hierarchical clustering methods to analyze ESTs can yield an optimal cDNA microarray design. If biomarker discovery is the goal of the microarray experiments, the average linkage method is more effective, while single linkage is more suitable if identification of physiological mechanisms is more of interest. This general design procedure is not limited to designing single-species cDNA microarrays for marine organisms, and it can equally be applied to multiple-species microarrays of any organisms with limited genomic information.