Gene Ontology Hierarchy Research Articles

A phylotranscriptomic analysis of gene family expansion and evolution in the largest order of pleurocarpous mosses (Hypnales, Bryophyta)

The pleurocarpous mosses (i.e., Hypnanae) are a species-rich group of land plants comprising about 6,000 species that share the development of female sex organs on short lateral branches, a derived trait within mosses. Many of the families within Hypnales, the largest order of pleurocarpous mosses, trace their origin to a rapid radiation less than 100 million years ago, just after the rise of the angiosperms. As a result, the phylogenetic resolution among families of Hypnales, necessary to test evolutionary hypotheses, has proven difficult using one or few loci. We present the first phylogenetic inference from high-throughput sequence data (transcriptome sequences) for pleurocarpous mosses. To test hypotheses of gene family evolution, we built a species tree of 21 pleurocarpous and six acrocarpous mosses using over one million sites from 659 orthologous genes. We used the species tree to investigate the genomic consequences of the shift to pleurocarpy and to identify whether patterns common to other plant radiations (gene family expansion, whole genome duplication, or changes in the molecular signatures of selection) could be observed. We found that roughly six percent of all gene families have expanded in the pleurocarpous mosses, relative to acrocarpous mosses. These gene families are enriched for several gene ontology (GO) terms, including interaction with other organisms. The increase in copy number coincident with the radiation of Hypnales suggests that a process such as whole genome duplication or a burst of small-scale duplications occurred during the diversification. In over 500 gene families we found evidence of a reduction in purifying selection. These gene families are enriched for several terms in the GO hierarchy related to “tRNA metabolic process.” Our results reveal candidate genes and pathways that may be associated with the transition to pleurocarpy, illustrating the utility of phylotranscriptomics for the study of molecular evolution in non-model species.

Predicting protein function via downward random walks on a gene ontology.

BackgroundHigh-throughput bio-techniques accumulate ever-increasing amount of genomic and proteomic data. These data are far from being functionally characterized, despite the advances in gene (or gene’s product proteins) functional annotations. Due to experimental techniques and to the research bias in biology, the regularly updated functional annotation databases, i.e., the Gene Ontology (GO), are far from being complete. Given the importance of protein functions for biological studies and drug design, proteins should be more comprehensively and precisely annotated.ResultsWe proposed downward Random Walks (dRW) to predict missing (or new) functions of partially annotated proteins. Particularly, we apply downward random walks with restart on the GO directed acyclic graph, along with the available functions of a protein, to estimate the probability of missing functions. To further boost the prediction accuracy, we extend dRW to dRW-kNN. dRW-kNN computes the semantic similarity between proteins based on the functional annotations of proteins; it then predicts functions based on the functions estimated by dRW, together with the functions associated with the k nearest proteins. Our proposed models can predict two kinds of missing functions: (i) the ones that are missing for a protein but associated with other proteins of interest; (ii) the ones that are not available for any protein of interest, but exist in the GO hierarchy. Experimental results on the proteins of Yeast and Human show that dRW and dRW-kNN can replenish functions more accurately than other related approaches, especially for sparse functions associated with no more than 10 proteins.ConclusionThe empirical study shows that the semantic similarity between GO terms and the ontology hierarchy play important roles in predicting protein function. The proposed dRW and dRW-kNN can serve as tools for replenishing functions of partially annotated proteins.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0713-y) contains supplementary material, which is available to authorized users.

Integrative meta-analysis of multiple gene expression profiles in acquired gemcitabine-resistant cancer cell lines to identify novel therapeutic biomarkers.

In molecular-targeted cancer therapy, acquired resistance to gemcitabine is a major clinical problem that reduces its effectiveness, resulting in recurrence and metastasis of cancers. In spite of great efforts to reveal the overall mechanism of acquired gemcitabine resistance, no definitive genetic factors have been identified that are absolutely responsible for the resistance process. Therefore, we performed a cross-platform meta-analysis of three publically available microarray datasets for cancer cell lines with acquired gemcitabine resistance, using the R-based RankProd algorithm, and were able to identify a total of 158 differentially expressed genes (DEGs; 76 up- and 82 down-regulated) that are potentially involved in acquired resistance to gemcitabine. Indeed, the top 20 up- and down-regulated DEGs are largely associated with a common process of carcinogenesis in many cells. For the top 50 up- and down-regulated DEGs, we conducted integrated analyses of a gene regulatory network, a gene co-expression network, and a protein-protein interaction network. The identified DEGs were functionally enriched via Gene Ontology hierarchy and Kyoto Encyclopedia of Genes and Genomes pathway analyses. By systemic combinational analysis of the three molecular networks, we could condense the total number of DEGs to final seven genes. Notably, GJA1, LEF1, and CCND2 were contained within the lists of the top 20 up- or down-regulated DEGs. Our study represents a comprehensive overview of the gene expression patterns associated with acquired gemcitabine resistance and theoretical support for further clinical therapeutic studies.

Systems pharmacology of mifepristone (RU486) reveals its 47 hub targets and network: comprehensive analysis and pharmacological focus on FAK-Src-Paxillin complex.

Mifepristone (RU486), a synthetic steroid compound used as an abortifacient drug, has received considerable attention to its anticancer activity recently. To explore the possibility of using mifepristone as a cancer metastasis chemopreventive, we performed a systems pharmacology analysis of mifepristone-related molecules in the present study. Data were collected by using Natural Language Processing (NLP) and 513 mifepristone-related genes were dug out and classified functionally using a gene ontology (GO) hierarchy, followed by KEGG pathway enrichment analysis. Potential signal pathways and targets involved in cancer were obtained by integrative network analysis. Total thirty-three proteins were involved in focal adhesion-the key signaling pathway associated with cancer metastasis. Molecular and cellular assays further demonstrated that mifepristone had the ability to prevent breast cancer cells from migration and interfere with their adhesion to endothelial cells. Moreover, mifepristone inhibited the expression of focal adhesion kinase (FAK), paxillin, and the formation of FAK/Src/Paxillin complex, which are correlated with cell adhesion and migration. This study set a good example to identify chemotherapeutic potential seamlessly from systems pharmacology to cellular pharmacology, and the revealed hub genes may be the promising targets for cancer metastasis chemoprevention.

Assessing protein-protein interactions based on the semantic similarity of interacting proteins.

The Gene Ontology (GO) has been used in estimating the semantic similarity of proteins since it has the largest and reliable vocabulary of gene products and characteristics. We developed a new method which can assess Protein-Protein Interactions (PPI) using the branching factor and information content of the common ancestor of interacting proteins in the GO hierarchy. We performed a comparative evaluation of the measure with other GO-based similarity measures and evaluation results showed that our method outperformed others in most GO domains.

Identifying significant associations of orthologous simple sequence repeats with gene ontologies.

Simple Sequence Repeats (SSRs), also known as microsatellites, regulate gene functions. SSR mutations in a disease gene may cause various genetic disorders. To identify putative functional SSRs, a web-based system, Gene Ontology SSR Hierarchy (GOSH), was developed to facilitate discovery of significant associations between SSRs and Gene Ontology (GO) terms. Using the GO hierarchy term structure, GOSH assists users with selecting functional or biological gene subsets. Significant SSR patterns are retrieved and identified via comprehensive overrepresentation analysis within a target gene subset and by comparing results with orthologous genes. Pattern relationships between different biological subsets or supersets can be observed by using the GO hierarchy structure directly. GOSH also supports GO searching through identified significant SSR patterns and all GO terms possessing such patterns are listed for consultation. GOSH is the first comprehensive and efficient online mining tool for discovering significant orthologous SSR patterns in GO terms and is available at http://gosh.cs.ntou.edu.tw/.

Construction of a cDNA Library from the Ephemeral Plant Olimarabidopsis pumila and Preliminary Analysis of Expressed Sequence Tags

Olimarabidopsis pumila is a close relative of the model plant Arabidopsis thaliana but, unlike A. thaliana, it is a salt-tolerant ephemeral plant that is widely distributed in semi-arid and semi-salinized regions of the Xinjiang region of China, thus providing an ideal candidate plant system for salt tolerance gene mining. A good-quality cDNA library was constructed using cap antibody to enrich full-length cDNA with the gateway technology allowing library construction without traditional methods of cloning by use of restriction enzymes. A preliminary analysis of expressed sequence tags (ESTs) was carried out. The titers of the primary and the normalized cDNA library were 1.6 x 10(6) cfu/mL and 6.7 x 10(6) cfu/mL, respectively. A total of 1093 clones were randomly selected from the normalized library for EST sequencing. By sequence analysis, 894 high-quality ESTs were generated and assembled into 736 unique sequences consisting of 72 contigs and 664 singletons. The resulting unigenes were categorized according to the gene ontology (GO) hierarchy. The potential roles of gene products associated with stress-related ESTs are discussed. The 736 unigenes were similar to A. thaliana, A. lyrata, or Thellungiella salsuginea. This research provides an overview of the mRNA expression profile and first-hand information of gene sequence expressed in young leaves of O. pumila.

Generation and analysis of expressed sequence tags fromthe salt-tolerant eelgrass species, Zostera marina

Zostera marina, a monocotyledonous angiosperm, is one of the most important seagrass species. To investigate the salt-tolerance mechanism and discover salt-tolerant genes in Z. marina, a cDNA library was constructed. Single-pass sequencing of the 5′ ends of 4 081 clones yielded 4 002 high quality expressed sequence tags (ESTs), which were assembled into 241 contigs and 1 673 singletons, representing 1 914 unigenes. The average length of the ESTs was 582 bp, with sizes ranging from 100–1 500 bp. Basic Local Alignment Search Tool (BLASTX) analysis revealed that 1 664 unigenes had significant homology to known genes in the National Center for Biotechnology Information (NCBI) non-redundant (nr) database (E-value⩽5–10). Among them, the two most abundant genes encoded metallothionein (157 ESTs) and chlorophyll a/b-binding protein (38 ESTs), accounting for 7.1% and 1.7% of the total ESTs, respectively. Using Kyoto Encyclopedia of Genes and Genomes (KEGG), 1 462 unigenes were assigned to 1 161 pathways (E-value⩽5–10). A total of 938 unigenes were assigned Gene Ontology (GO) terms based on the GO hierarchy analysis, and InterProScan searches recognized 1 003 InterPro families. Three genes for metallothionein in Z. marina that belonged to Class II was identified. Results of this study will improve understanding of the molecular mechanisms of saline tolerance in Z. marina.

Integration of genome-wide association studies with biological knowledge identifies six novel genes related to kidney function

In conducting genome-wide association studies (GWAS), analytical approaches leveraging biological information may further understanding of the pathophysiology of clinical traits. To discover novel associations with estimated glomerular filtration rate (eGFR), a measure of kidney function, we developed a strategy for integrating prior biological knowledge into the existing GWAS data for eGFR from the CKDGen Consortium. Our strategy focuses on single nucleotide polymorphism (SNPs) in genes that are connected by functional evidence, determined by literature mining and gene ontology (GO) hierarchies, to genes near previously validated eGFR associations. It then requires association thresholds consistent with multiple testing, and finally evaluates novel candidates by independent replication. Among the samples of European ancestry, we identified a genome-wide significant SNP in FBXL20 (P = 5.6 × 10(-9)) in meta-analysis of all available data, and additional SNPs at the INHBC, LRP2, PLEKHA1, SLC3A2 and SLC7A6 genes meeting multiple-testing corrected significance for replication and overall P-values of 4.5 × 10(-4)-2.2 × 10(-7). Neither the novel PLEKHA1 nor FBXL20 associations, both further supported by association with eGFR among African Americans and with transcript abundance, would have been implicated by eGFR candidate gene approaches. LRP2, encoding the megalin receptor, was identified through connection with the previously known eGFR gene DAB2 and extends understanding of the megalin system in kidney function. These findings highlight integration of existing genome-wide association data with independent biological knowledge to uncover novel candidate eGFR associations, including candidates lacking known connections to kidney-specific pathways. The strategy may also be applicable to other clinical phenotypes, although more testing will be needed to assess its potential for discovery in general.

Differential Display of Abundantly Expressed Genes of Trichoderma harzianum During Colonization of Tomato-Germinating Seeds and Roots

The identification of Trichoderma genes whose expression is altered during early stages of interaction with developing roots of germinated seeds is an important step toward understanding the rhizosphere competency of Trichoderma spp. The potential of 13 Trichoderma strains to colonize tomato root and promote plant growth has been evaluated. All used strains successfully propagated in spermosphere and continued their growth in rhizoplane simultaneously root enlargement while the strains T6 and T7 were the most abundant in the apical segment of roots. Root colonization in most strains associated with promoting the roots and shoots growth while they significantly increased up to 43 and 40% roots and shoots dry weights, respectively. Differential display reverse transcriptase-PCR (DDRT-PCR) has been developed to detect differentially expressed genes in the previously selected strain, Trichoderma harzianum T7, during colonization stages of tomato-germinating seeds and roots. Amplified DDRT-PCR products were analyzed on gel agarose and 62 differential bands excised, purified, cloned, and sequenced. Obtained ESTs were submit-queried to NCBI database by BLASTx search and gene ontology hierarchy. Most of transcripts (29 EST) corresponds to known and hypothetical proteins such as secretion-related small GTPase, 40S ribosomal protein S3a, 3-hydroxybutyryl-CoA dehydrogenase, DNA repair protein rad50, lipid phosphate phosphatase-related protein type 3, nuclear essential protein, phospholipase A2, fatty acid desaturase, nuclear pore complex subunit Nup133, ubiquitin-activating enzyme, and 60S ribosomal protein L40. Also, 13 of these sequences showed no homology (E>0.05) with public databases and considered as novel genes. Some of these ESTs corresponded to genes encodes enzymes potentially involved in nutritional support of microorganisms which have obvious importance in the establishment of Trichoderma in spermosphere and rhizosphere, via potentially functioning in acquisition of nutrients from energy-rich carbon compounds leaked from the germinating seeds and roots.

Identification of drought responsive proteins using gene ontology hierarchy.

The availability of the complete genome sequences has facilitated access to essential information to identify proteins. The determination of Arabidopsis genome sequence has had a great impact to annotate data. The genome sequencing of Sorghum bicolor has been only recently completed and hither to the global response to abiotic stresses in this important crop remains largely unexplored. We used 2-D gel electrophoresis based proteomic approach refined with MALDI-TOF to analyze drought-stress response proteins in sorghum. Major changes in protein complement of sorghum were observed in hydroponic cultures at 96 hours under drought stress. Six most highly expressed proteins were excised for functional identification. Here, we developed a method to obtain functional distances between GO terms and analyzed distance values to allocate shortest path (SP) in GO hierarchy. The shortest paths for expressed proteins were noted for most informative common ancestor (MICA) terms, viz. binding, catalytic activity and primary metabolic process. We observed the expressed proteins belonged to the functional group of signal transduction mechanisms, carbohydrate transport and metabolism. These identified functions of proteins suggest a different mechanism of drought-stress tolerant in sorghum. The novel approach applied in this study may have great importance in further identifying proteins involved in abiotic and biotic stress conditions in crops.

Identification of salt treated proteins in sorghum using gene ontology linkage

Sorghum bicolor (L.) is an important crop of arid and semi arid zones with most of its varieties tolerant to drought, heat and salt stress. Functional identification of many salt tolerant proteins has been reported in Arabidopsis, rice and other plants, however only little functional information has been predicted in sorghum till date. A 2-D gel electrophoresis based proteomic approach with MALDI-TOF mass spectrometer was utilized to analyze the salt stress response of sorghum. Major changes in protein complement were observed at 200mM NaCl in hydroponic culture after 96h of salt-stress. Highly expressed five proteins were excised for functional identification. We developed shortest path (SP) analysis based method on Gene Ontology (GO) hierarchy using sum of GO-term's semantic similarities. In this study, we observed that majority of expressed proteins belonged to the functional category of energy production and conversion, signal transduction mechanisms and ribosome maturation. These identified functions suggest a distinct mechanism of salt-stress adaptation in sorghum plant. The proposed method in this paper potentially has great importance to further understanding of newly identified proteins that can help in plant development.

PPISearchEngine: gene ontology-based search for protein–protein interactions

This paper presents a new search engine called PPISearchEngine which finds protein–protein interactions (PPIs) using the gene ontology (GO) and the biological relations of proteins. For efficient retrieval of PPIs, each GO term is assigned a prime number and the relation between the terms is represented by the product of prime numbers. This representation is hidden from users but facilitates the search for the interactions of a query protein by unique prime factorisation of the number that represents the query protein. For a query protein, PPISearchEngine considers not only the GO term associated with the query protein but also the GO terms at the lower level than the GO term in the GO hierarchy, and finds all the interactions of the query protein which satisfy the search condition. In contrast, the standard keyword-matching or ID-matching search method cannot find the interactions of a protein unless the interactions involve a protein with explicit annotations. To the best of our knowledge, this search engine is the first method that can process queries like ‘for protein p with GO , find p's interaction partners with GO ’. PPISearchEngine is freely available to academics at http://search.hpid.org/.

Characterization of resistance to pine wood nematode infection in Pinus thunbergiiusing suppression subtractive hybridization

BackgroundPine wilt disease is caused by the pine wood nematode, Bursaphelenchus xylophilus, which threatens pine forests and forest ecosystems worldwide and causes serious economic losses. In the 40 years since the pathogen was identified, the physiological changes occurring as the disease progresses have been characterized using anatomical and biochemical methods, and resistant trees have been selected via breeding programs. However, no studies have assessed the molecular genetics, e.g. transcriptional changes, associated with infection-induced physiological changes in resistant or susceptible trees.ResultsWe constructed seven subtractive suppression hybridization (SSH) cDNA libraries using time-course sampling of trees inoculated with pine wood nematode at 1, 3, or 7 days post-inoculation (dpi) in susceptible trees and at 1, 3, 7, or 14 dpi in resistant trees. A total of 3,299 sequences was obtained from these cDNA libraries, including from 138 to 315 non-redundant sequences in susceptible SSH libraries and from 351 to 435 in resistant SSH libraries. Using Gene Ontology hierarchy, those non-redundant sequences were classified into 15 subcategories of the biological process Gene Ontology category and 17 subcategories of the molecular function category. The transcriptional components revealed by the Gene Ontology classification clearly differed between resistant and susceptible libraries. Some transcripts were discriminative: expression of antimicrobial peptide and putative pathogenesis-related genes (e.g., PR-1b, 2, 3, 4, 5, 6) was much higher in susceptible trees than in resistant trees at every time point, whereas expression of PR-9, PR-10, and cell wall-related genes (e.g., for hydroxyproline-rich glycoprotein precursor and extensin) was higher in resistant trees than in susceptible trees at 7 and 14 dpi.ConclusionsFollowing inoculation with pine wood nematode, there were marked differences between resistant and susceptible trees in transcript diversity and the timing and level of transcripts expressed in common; in particular, expression of stress response and defense genes differed. This study provided new insight into the differences in the physiological changes between resistant and susceptible trees that have been observed in anatomical and biochemical studies.

Generation and Analysis of Expressed Sequence Tags fromChimonanthus praecox(Wintersweet) Flowers for Discovering Stress-Responsive and Floral Development-Related Genes

A complementary DNA library was constructed from the flowers of Chimonanthus praecox, an ornamental perennial shrub blossoming in winter in China. Eight hundred sixty-seven high-quality expressed sequence tag sequences with an average read length of 673.8 bp were acquired. A nonredundant set of 479 unigenes, including 94 contigs and 385 singletons, was identified after the expressed sequence tags were clustered and assembled. BLAST analysis against the nonredundant protein database and nonredundant nucleotide database revealed that 405 unigenes shared significant homology with known genes. The homologous unigenes were categorized according to Gene Ontology hierarchies (biological, cellular, and molecular). By BLAST analysis and Gene Ontology annotation, 95 unigenes involved in stress and defense and 19 unigenes related to floral development were identified based on existing knowledge. Twelve genes, of which 9 were annotated as “cold response,” were examined by real-time RT-PCR to understand the changes in expression patterns under cold stress and to validate the findings. Fourteen genes, including 11 genes related to floral development, were also detected by real-time RT-PCR to validate the expression patterns in the blooming process and in different tissues. This study provides a useful basis for the genomic analysis of C. praecox.

Targeted mining of drought stress-responsive genes from EST resources in Cleistogenes songorica

Cleistogenes songorica is an important perennial grass found in the pastoral steppe of Inner Mongolia. C. songorica flourishes in drought prone environments, and therefore provides an ideal candidate plant system for the identification of drought-tolerance conferring genes. We constructed cDNA libraries from leaves and roots of drought-stressed C. songorica seedlings. Expressed sequence tag (EST) sequencing of 5664 random cDNA clones produced 3579 high quality, trimmed sequences. The average read length of trimmed ESTs was 613 bp. Clustering and assembly identified a non-redundant set of 1499 contigs, including 805 singleton unigenes and 694 multi-member unigenes. The resulting unigenes were functionally categorized according to the Gene Ontology (GO) hierarchy using the in house Bioinformatic Advanced Scientific Computing (BASC) annotation pipeline. Among the total 2.2 Mbp of EST sequence data, 161 putative SSRs were found, a frequency similar to that previously observed in oat and Arabidopsis ESTs. Sixty-three unigenes were functionally annotated as being stress responsive, of which 22 were similar to genes implicated in drought stress response. Using quantitative real time RT-PCR, transcripts of 13 of these 22 genes were shown to be at least three fold more, or less abundant in drought-stressed leaves or roots, with 8 increased and 5 decreased in relative transcript abundance. The C. songorica EST and cDNA collections generated in this study are a valuable resource for microarray-based expression profiling, and functional genomics in order to elucidate their role, and to understand the underlying mechanisms of drought-tolerance in C. songorica.

Generation and analysis of expressed sequence tags from the bone marrow of Chinese Sika deer

Sika deer is one of the best-known and highly valued animals of China. Despite its economic, cultural, and biological importance, there has not been a large-scale sequencing project for Sika deer to date. With the ultimate goal of sequencing the complete genome of this organism, we first established a bone marrow cDNA library for Sika deer and generated a total of 2,025 reads. After processing the sequences, 2,017 high-quality expressed sequence tags (ESTs) were obtained. These ESTs were assembled into 1,157 unigenes, including 238 contigs and 919 singletons. Comparative analyses indicated that 888 (76.75%) of the unigenes had significant matches to sequences in the non-redundant protein database, In addition to highly expressed genes, such as stearoyl-CoA desaturase, cytochrome c oxidase, adipocyte-type fatty acid-binding protein, adiponectin and thymosin beta-4, we also obtained vascular endothelial growth factor-A and heparin-binding growth-associated molecule, both of which are of great importance for angiogenesis research. There were 244 (21.09%) unigenes with no significant match to any sequence in current protein or nucleotide databases, and these sequences may represent genes with unknown function in Sika deer. Open reading frame analysis of the sequences was performed using the getorf program. In addition, the sequences were functionally classified using the gene ontology hierarchy, clusters of orthologous groups of proteins and Kyoto encyclopedia of genes and genomes databases. Analysis of ESTs described in this paper provides an important resource for the transcriptome exploration of Sika deer, and will also facilitate further studies on functional genomics, gene discovery and genome annotation of Sika deer.

Analysis of Expressed Sequence Tags from the Wood-Decaying Fungus Fomitopsis palustris and Identification of Potential Genes Involved in the Decay Process

Fomitopsis palustris, a brown-rot basidiomycete, causes the most destructive type of decay in wooden structures. In spite of its great economic importance, very little information is available at the molecular level regarding its complex decay process. To address this, we generated over 3,000 expressed sequence tags (ESTs) from a cDNA library constructed from F. palustris. Clustering of 3,095 high-quality ESTs resulted in a set of 1,403 putative unigenes comprising 485 contigs and 918 singlets. Homology searches based on BlastX analysis revealed that 78% of the F. palustris unigenes had a significant match to proteins deposited in the nonredundant databases. A subset of F. palustris unigenes showed similarity to the carbohydrateactive enzymes (CAZymes), including a range of glycosyl hydrolase (GH) family proteins. Some of these CAZymeencoded genes were previously undescribed for F. palustris but predicted to have potential roles in biodegradation of wood. Among them, we identified and characterized a gene (FpCel45A) encoding the GH family 45 endoglucanase. Moreover, we also provided functional classification of 473 (34%) of F. palustris unigenes using the Gene Ontology hierarchy. The annotated EST data sets and related analysis may be useful in providing an initial insight into the genetic background of F. palustris.

Abstract 36: Molecular profiling of prostate cancer xenografts and human specimens reveals overexpression of UBE2C/UBCH10 and activation of aurora kinases in poorly differentiated neuroendocrine carcinoma of the prostate

Abstract Small cell and large cell neuroendocrine carcinoma belong to the spectrum of poorly differentiated neuroendocrine carcinomas (NECa) of the prostate, characterized by a distinct clinical course. Often unrecognized, they can emerge either de novo or in the castrate-resistant progression of typical acinar adenocarcinoma (AdCa) and are found in 10-20% of men who die of the disease. Although sensitive to chemotherapy, responses are short lived and prognosis is dismal. An understanding of the biology underlying NECa and identification of novel therapeutic targets for this variant are urgently needed. In this study we used xenograft models and tissue specimens obtained from castrate-resistant prostate carcinomas (CRPC) to gain insight into the pathways implicated in the development of the highly aggressive NECa. Xenograft models with features of typical AdCa (n=3) and NECa (n=4) were subjected to gene-expression profiling using Affymetrix HGU133 Plus 2.0. One-way ANOVA was applied to identify differentially expressed probes (DEPs). The Web-based Gene Ontology (GO) Tree Machine was used to examine GO hierarchies. Selected genes were validated by quantitative real-time PCR (qRT-PCR) and Western blotting in the xenografts and by immunohistochemistry in the xenografts, and the tumor specimens of 62 patients with CRPC. Unsupervised hierarchal clustering of the raw data classified the xenografts correctly according to their morphology. Using stringent FDR 0.05, we identified 140 DEPs (0.3%) between the two groups. GO analysis revealed significant enrichment in the “M phase of mitotic cell cycle” biological process subtree (adjP=7.04e-08). Amongst the genes differentially expressed between the two groups, UBE2C, (E2 ubiquitin-conjugating enzyme UBCH10), a member of the anaphase promoting complex (APC), was significantly upregulated in NECa relative to AdCa by qRT-PCR (P=0.003) and immunohistochemistry (P=0.007) in xenografts, and by immunohistochemistry in the CRPC specimens (P<0.001). Similarly, Aurora kinases A and B, proteins with key roles during the G2/M phase of cell cycle, were significantly activated in NECa compared to AdCa xenografts as shown by Western blotting using phospho-Aurora A (Thr288)/Aurora B (Thr232) specific antibodies. In contrast, Cyclin D1, a G1-S phase regulator, was significantly downregulated in NECa xenografts relative to AdCa at the mRNA (P=0.049) and protein level (P=0.003). Cyclin D1 downregulation was also confirmed in the CRPC specimens with NECa relative to those with AdCa morphology (P<0.001). Our findings of UBE2C upregulation and Aurora A and B activation in prostate NECa, suggest a role for G2/M and APC dysregulation in the pathogenesis of NECa and provide the rational to pursue them as therapeutic targets in this prostate cancer variant. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 36. doi:10.1158/1538-7445.AM2011-36

Recovering key biological constituents through sparse representation of gene expression

Large-scale RNA expression measurements are generating enormous quantities of data. During the last two decades, many methods were developed for extracting insights regarding the interrelationships between genes from such data. The mathematical and computational perspectives that underlie these methods are usually algebraic or probabilistic. Here, we introduce an unexplored geometric view point where expression levels of genes in multiple experiments are interpreted as vectors in a high-dimensional space. Specifically, we find, for the expression profile of each particular gene, its approximation as a linear combination of profiles of a few other genes. This method is inspired by recent developments in the realm of compressed sensing in the machine learning domain. To demonstrate the power of our approach in extracting valuable information from the expression data, we independently applied it to large-scale experiments carried out on the yeast and malaria parasite whole transcriptomes. The parameters extracted from the sparse reconstruction of the expression profiles, when fed to a supervised learning platform, were used to successfully predict the relationships between genes throughout the Gene Ontology hierarchy and protein-protein interaction map. Extensive assessment of the biological results shows high accuracy in both recovering known predictions and in yielding accurate predictions missing from the current databases. We suggest that the geometrical approach presented here is suitable for a broad range of high-dimensional experimental data.