Learning with Kernels on Graphs: DAG-based kernels, data streams and RNA function prediction.

In many application domains data can be naturally represented as graphs. When the application of analytical solutions for a given problem is unfeasible, machine learning techniques could be a viable way to solve the problem. Classical machine learning techniques are defined for data represented in a vectorial form. Recently some of them have been extended to deal directly with structured data. Among those techniques, kernel methods have shown promising results both from the computational complexity and the predictive performance point of view. Kernel methods allow to avoid an explicit mapping in a vectorial form relying on kernel functions, which informally are functions calculating a similarity measure between two entities. However, the definition of good kernels for graphs is a challenging problem because of the difficulty to find a good tradeoff between computational complexity and expressiveness. Another problem we face is learning on data streams, where a potentially unbounded sequence of data is generated by some sources. There are three main contributions in this thesis. The first contribution is the definition of a new family of kernels for graphs based on Directed Acyclic Graphs (DAGs). We analyzed two kernels from this family, achieving state-of-the-art results from both the computational and the classification point of view on real-world datasets. The second contribution consists in making the application of learning algorithms for streams of graphs feasible. Moreover,we defined a principled way for the memory management. The third contribution is the application of machine learning techniques for structured data to non-coding RNA function prediction. In this setting, the secondary structure is thought to carry relevant information. However, existing methods considering the secondary structure have prohibitively high computational complexity. We propose to apply kernel methods on this domain, obtaining state-of-the-art results.

Novel Primate-Specific Genes, RMEL 1, 2 and 3, with Highly Restricted Expression in Melanoma, Assessed by New Data Mining Tool

Melanoma is a highly aggressive and therapy resistant tumor for which the identification of specific markers and therapeutic targets is highly desirable. We describe here the development and use of a bioinformatic pipeline tool, made publicly available under the name of EST2TSE, for the in silico detection of candidate genes with tissue-specific expression. Using this tool we mined the human EST (Expressed Sequence Tag) database for sequences derived exclusively from melanoma. We found 29 UniGene clusters of multiple ESTs with the potential to predict novel genes with melanoma-specific expression. Using a diverse panel of human tissues and cell lines, we validated the expression of a subset of three previously uncharacterized genes (clusters Hs.295012, Hs.518391, and Hs.559350) to be highly restricted to melanoma/melanocytes and named them RMEL1, 2 and 3, respectively. Expression analysis in nevi, primary melanomas, and metastatic melanomas revealed RMEL1 as a novel melanocytic lineage-specific gene up-regulated during melanoma development. RMEL2 expression was restricted to melanoma tissues and glioblastoma. RMEL3 showed strong up-regulation in nevi and was lost in metastatic tumors. Interestingly, we found correlations of RMEL2 and RMEL3 expression with improved patient outcome, suggesting tumor and/or metastasis suppressor functions for these genes. The three genes are composed of multiple exons and map to 2q12.2, 1q25.3, and 5q11.2, respectively. They are well conserved throughout primates, but not other genomes, and were predicted as having no coding potential, although primate-conserved and human-specific short ORFs could be found. Hairpin RNA secondary structures were also predicted. Concluding, this work offers new melanoma-specific genes for future validation as prognostic markers or as targets for the development of therapeutic strategies to treat melanoma.

A score system for quality evaluation of RNA sequence tags: an improvement for gene expression profiling

Background: High-throughput molecular approaches for gene expression profiling, such as Serial Analysis of Gene Expression (SAGE), Massively Parallel Signature Sequencing (MPSS) or Sequencing-by-Synthesis (SBS) represent powerful techniques that provide global transcription profiles of different cell types through sequencing of short fragments of transcripts, denominated sequence tags. These techniques have improved our understanding about the relationships between these expression profiles and cellular phenotypes. Despite this, more reliable datasets are still necessary. In this work, we present a web-based tool named S3T: Score System for Sequence Tags, to index sequenced tags in accordance with their reliability. This is made through a series of evaluations based on a defined rule set. S3T allows the identification/selection of tags, considered more reliable for further gene expression analysis. Results: This methodology was applied to a public SAGE dataset. In order to compare data before and after filtering, a hierarchical clustering analysis was performed in samples from the same type of tissue, in distinct biological conditions, using these two datasets. Our results provide evidences suggesting that it is possible to find more congruous clusters after using S3T scoring system. Conclusion: These results substantiate the proposed application to generate more reliable data. This is a significant contribution for determination of global gene expression profiles. The library analysis with S3T is freely available at http://gdm.fmrp.usp.br/s3t/.S3T source code and datasets can also be downloaded from the aforementioned website.

The taxonomic status of the endangered thin-spined porcupine, Chaetomys subspinosus (Olfers, 1818), based on molecular and karyologic data

Background: The thin-spined porcupine, also known as the bristle-spined rat, Chaetomys subspinosus (Olfers, 1818), the only member of its genus, figures among Brazilian endangered species. In addition to being threatened, it is poorly known, and even its taxonomic status at the family level has long been controversial. The genus Chaetomys was originally regarded as a porcupine in the family Erethizontidae, but some authors classified it as a spiny-rat in the family Echimyidae. Although the dispute seems to be settled in favor of the erethizontid advocates, further discussion of its affinities should be based on a phylogenetic framework. In the present study, we used nucleotide-sequence data from the complete mitochondrial cytochrome b gene and karyotypic information to address this issue. Our molecular analyses included one individual of Chaetomys subspinosus from the state of Bahia in northeastern Brazil, and other hystricognaths. Results: All topologies recovered in our molecular phylogenetic analyses strongly supported Chaetomys subspinosus as a sister clade of the erethizontids. Cytogenetically, Chaetomys subspinosus showed 2n = 52 and FN = 76. Although the sexual pair could not be identified, we assumed that the X chromosome is biarmed. The karyotype included 13 large to medium metacentric and submetacentric chromosome pairs, one small subtelocentric pair, and 12 small acrocentric pairs. The subtelocentric pair 14 had a terminal secondary constriction in the short arm, corresponding to the nucleolar organizer region (Ag-NOR), similar to the erethizontid Sphiggurus villosus, 2n = 42 and FN = 76, and different from the echimyids, in which the secondary constriction is interstitial. Conclusion: Both molecular phylogenies and karyotypical evidence indicated that Chaetomys is closely related to the Erethizontidae rather than to the Echimyidae, although in a basal position relative to the rest of the Erethizontidae. The high levels of molecular and morphological divergence suggest that Chaetomys belongs to an early radiation of the Erethizontidae that may have occurred in the Early Miocene, and should be assigned to its own subfamily, the Chaetomyinae.

Mitochondrial genome data alone are not enough to unambiguously resolve the relationships of Entognatha, Insecta and Crustacea sensu lato (Arthropoda)

An analysis of the relationships of the major arthropod groups Was undertaken using mitochondrial genome data to examine the hypotheses that Hexapoda is polyphyletic and that Collembola is more closely related to branchiopod crustaceans than insects. We sought to examine the sensitivity of this relationship to outgroup choice, data treatment. gene choice and optimality criteria used in the phylogenetic analysis of mitochondrial genome data. Additionally we sequenced the mitochondrial genome of ail archaeognathan, Nesomachilis australica. to improve taxon selection in the apterygote insects, a group poorly represented in previous mitochondrial phylogenies. The sister group of the Collembola was rarely resolved in our analyses with a significant level of support. The use of different outgroups (myriapods, nematodes, or annelids + mollusks) resulted in many different placements of Collembola. The way in which the dataset was coded for analysis (DNA, DNA with the exclusion of third codon position and as amino acids) also had marked affects on tree topology. We found that nodal Support was spread evenly throughout the 13 mitochondrial genes and the exclusion of genes resulted in significantly less resolution in the inferred trees. Optimality criteria had a much lesser effect on topology than the preceding factors; parsimony and Bayesian trees for a given data set and treatment were quite similar. We therefore conclude that the relationships of the extant arthropod groups as inferred by mitochondrial genomes are highly vulnerable to outgroup choice, data treatment and gene choice, and no consistent alternative hypothesis of Collembola's relationships is supported. Pending the resolution of these identified problems with the application of mitogenomic data to basal arthropod relationships, it is difficult to justify the rejection of hexapod monophyly, which is well supported on morphological grounds. (c) The Willi Hennig Society 2004.

NOVEL 65 kDa RNA-BINDING PROTEIN IN SQUID PRESYNAPTIC TERMINALS

A polyclonal antibody (C4), raised against the head domain of chicken myosin Va, reacted strongly towards a 65 kDa polypeptide (p65) on Western blots of extracts from squid optic lobes but did not recognize the heavy chain of squid myosin V. This peptide was not recognized by other myosin Va antibodies, nor by an antibody specific for squid myosin V. In an attempt to identify it, p65 was purified from optic lobes of Loligo plei by cationic exchange and reverse phase chromatography. Several peptide sequences were obtained by mass spectroscopy from p65 cut from sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) gels. BLAST analysis and partial matching with expressed sequence tags (ESTs) from a Loligo pealei data bank indicated that p65 contains consensus signatures for the heterogeneous nuclear ribonucleoprotein (hnRNP) A/B family of RNA-binding proteins. Centrifugation of post mitochondrial extracts from optic lobes on sucrose gradients after treatment with RNase gave biochemical evidence that p65 associates with cytoplasmic RNP complexes in an RNA-dependent manner. Immunohistochemistry and immunofluorescence studies using the C4 antibody showed partial co-labeling with an antibody against squid synaptotagmin in bands within the outer plexiform layer of the optic lobes and at the presynaptic zone of the stellate ganglion. Also, punctate labeling by the C4 antibody was observed within isolated optic lobe synaptosomes. The data indicate that p65 is a novel RNA-binding protein located to the presynaptic terminal within squid neurons and may have a role in synaptic localization of RNA and its translation or processing. (C) 2010 IBRO. Published by Elsevier Ltd. All rights reserved.

The ratio of messenger RNA levels of receptor activator of nuclear factor kappa B ligand to osteoprotegerin correlates with bone remodeling indices in normal human cancellous bone but not in osteoarthritis

skeletal disease. Bone remodeling is initiated by osteoclastic resorption followed by osteoblastic formation of new bone. Receptor activator of nuclear factor KB ligand (RANKL) is a newly described regulator of osteoclast formation and function, the activity of which appears to be a balance between interaction with its receptor RANK and with an antagonist binding protein osteoprotegerin (OPG). Therefore, we have examined the relationship between the expression of RANKL, RANK, and OPG and indices of bone structure and turnover in human cancellous bone from the proximal femur. Bone samples were obtained from individuals with osteoarthritis (OA) at joint replacement surgery and from autopsy controls. Histomorphometric analysis of these samples showed that eroded surface (ES/BS) and osteoid surface (OS/BS) were positively associated in both control (p < 0.001) and OA (p < 0.02), indicating that the processes of bone resorption and bone formation remain coupled in OA, as they are in controls. RANKL, OPG, and RANK messenger RNA, (mRNA) were abundant in human cancellous bone, with significant differences between control and OA individuals. In coplotting the molecular and histomorphometric data, strong associations were found between the ratio of RANKL/OPG mRNA and the indices of bone turnover (RANKL/OPG vs. ES/BS: r = 0.93, p < 0.001; RANKL/OPG vs. OS/BS: r = 0.80, p < 0.001). These relationships were not evident in trabecular bone from severe OA, suggesting that bone turnover may be regulated differently in this disease. We propose that the effective concentration of RANKL is related causally to bone turnover.

Heterogeneous nuclear ribonucleoprotein A3, a novel RNA trafficking response element-binding protein

The cis-acting response element, A2RE, which is sufficient for cytoplasmic mRNA trafficking in oligodendrocytes, binds a small group of rat brain proteins. Predominant among these is heterogeneous nuclear ribonucleoprotein (hnRNP) A2, a trans-acting factor for cytoplasmic trafficking of RNAs bearing A2RE-like sequences. We have now identified the other A2RE-binding proteins as hnRNP A1/A1(B), hnRNP B1, and four isoforms of hnRNP A3. The rat and human hnRNP A3 cDNAs have been sequenced, revealing the existence of alternatively spliced mRNAs. In Western blotting, 38-, 39-, 41 -, and 41.5-kDa components were all recognized by antibodies against a peptide in the glycine-rich region of hnRNP A3, but only the 41- and 41.5-kDa bands bound antibodies to a 15-residue N-terminal peptide encoded by an alternatively spliced part of exon 1. The identities of these four proteins were verified by Edman sequencing and mass spectral analysis of tryptic fragments generated from electrophoretically separated bands. Sequence-specific binding of bacterially expressed hnRNP A3 to A2RE has been demonstrated by biosensor and UV cross-linking electrophoretic mobility shift assays. Mutational analysis and confocal microscopy data support the hypothesis that the hnRNP A3 isoforms have a role in cytoplasmic trafficking of RNA.

Molecular and functional analyses of Kunjin virus infectious cDNA clones demonstrate the essential roles for NS2A in virus assembly and for a nonconservative residue in NS3 in RNA replication

A number of full-length cDNA clones of Kunjin virus (KUN) were previously prepared; it was shown that two of them, pAKUN and FLSDX, differed in specific infectivities of corresponding in vitro transcribed RNAs by similar to100,000-fold (A. A. Khromykh et al., J. Virol. 72:7270-7279, 1998). In this study, we analyzed a possible genetic determinant(s) of the observed differences in infectivity initially by sequencing the entire cDNAs of both clones and comparing them with the published sequence of the parental KUN strain MRM61C. We found six common amino acid residues in both cDNA clones that were different from those in the published MRM61C sequence but were similar to those in the published sequences of other flaviviruses from the same subgroup. pAKUN clone had four additional codon changes, i.e., Ile59 to Asn and Arg175 to Lys in NS2A and Tyr518 to His and Ser557 to Pro in NS3. Three of these substitutions except the previously shown marker mutation, Arg175 to Lys in NS2A, reverted to the wild-type sequence in the virus eventually recovered from pAKUN RNA-transfected BHK cells, demonstrating the functional importance of these residues in viral replication and/or viral assembly. Exchange of corresponding DNA fragments between pAKUN and FLSDX clones and site-directed mutagenesis revealed that the Tyr518-to-His mutation in NS3 was responsible for an similar to5-fold decrease in specific infectivity of transcribed RNA, while the Ile59-to-Asn mutation in NS2A completely blocked virus production. Correction of the Asn59 in pAKUN NS2A to the wild-type lie residue resulted in complete restoration of RNA infectivity. Replication of KUN replicon RNA with an Ile59-to-Asn substitution in NS2A and with a Ser557-to-Pro substitution in NS3 was not affected, while the Tyr518-to-His substitution in NS3 led to severe inhibition of RNA replication. The impaired function of the mutated NS2A in production of infectious virus was complemented in trans by the helper wild-type NS2A produced from the KUN replicon RNA. However, replicon RNA with mutated NS2A could not be packaged in trans by the KUN structural proteins. The data demonstrated essential roles for the KUN nonstructural protein NS2A in virus assembly and for NS3 in RNA replication and identified specific single-amino-acid residues involved in these functions.

DNA vaccine coding for the full-length infectious Kunjin virus RNA protects mice against the New York strain of West Nile virus

A plasmid DNA directing transcription of the infectious full-length RNA genome of Kunjin (KUN) virus in vivo from a mammalian expression promoter was used to vaccinate mice intramuscularly. The KUN viral cDNA encoded in the plasmid contained the mutation in the NS1 protein (Pro-250 to Leu) previously shown to attenuate KUN virus in weanling mice. KUN virus was isolated from the blood of immunized mice 3-4 days after DNA inoculation, demonstrating that infectious RNA was being transcribed in vivo; however, no symptoms of virus-induced disease were observed. By 19 days postimmunization, neutralizing antibody was detected in the serum of immunized animals. On challenge with lethal doses of the virulent New York strain of West Nile (WN) or wild-type KUN virus intracerebrally or intraperitoneally, mice immunized with as little as 0.1-1 mug of KUN plasmid DNA were solidly protected against disease. This finding correlated with neutralization data in vitro showing that serum from KUN DNA-immunized mice neutralized KUN and WN,viruses with similar efficiencies. The results demonstrate that delivery of an attenuated but replicating KUN virus via a plasmid DNA vector may provide an effective vaccination strategy against virulent strains of WN virus.

The expression of tubulin cofactor A (TBCA) is regulated by a noncoding antisense Tbca RNA during testis maturation

Abstract - Recently, long noncoding RNAs have emerged as pivotal molecules for the regulation of coding genes' expression. These molecules might result from antisense transcription of functional genes originating natural antisense transcripts (NATs) or from transcriptional active pseudogenes. TBCA interacts with β-tubulin and is involved in the folding and dimerization of new tubulin heterodimers, the building blocks of microtubules. Methodology/Principal findings: We found that the mouse genome contains two structurally distinct Tbca genes located in chromosomes 13 (Tbca13) and 16 (Tbca16). Interestingly, the two Tbca genes albeit ubiquitously expressed, present differential expression during mouse testis maturation. In fact, as testis maturation progresses Tbca13 mRNA levels increase progressively, while Tbca16 mRNA levels decrease. This suggests a regulatory mechanism between the two genes and prompted us to investigate the presence of the two proteins. However, using tandem mass spectrometry we were unable to identify the TBCA16 protein in testis extracts even in those corresponding to the maturation step with the highest levels of Tbca16 transcripts. These puzzling results led us to re-analyze the expression of Tbca16. We then detected that Tbca16 transcription produces sense and natural antisense transcripts. Strikingly, the specific depletion by RNAi of these transcripts leads to an increase of Tbca13 transcript levels in a mouse spermatocyte cell line. Conclusions/Significance: Our results demonstrate that Tbca13 mRNA levels are post-transcriptionally regulated by the sense and natural antisense Tbca16 mRNA levels. We propose that this regulatory mechanism operates during spermatogenesis, a process that involves microtubule rearrangements, the assembly of specific microtubule structures and requires critical TBCA levels.

Quantification of HIV-1 viral RNA in the blood in needles used for venous puncture in HIV-infected individuals

INTRODUCTION: Occupational HIV infection among healthcare workers is an important issue in exposures involving blood and body fluids. There are few data in the literature regarding the potential and the duration of infectivity of HIV type 1 (HIV-1) in contaminated material under adverse conditions. METHODS: We quantified HIV-1 viral RNA in 25×8mm calibre hollow-bore needles, after punctures, in 25 HIV-1-infected patients selected during the sample collection. All of the patients selected were between the ages of 18 and 55. Five samples were collected from 16 patients: one sample for the immediate quantification of HIV-1 RNA in the plasma and blood samples from the interior of 4 needles to be analyzed at 0h, 6h, 24h, and 72h after collection. In nine patients, another test was carried out in the blood from one additional needle, in which HIV-1 RNA was assessed 168h after blood collection. The method used to assess HIV-1 RNA was nucleic acid sequence-based amplification. RESULTS: Up to 7 days after collection, HIV-1 RNA was detected in all of the needles. The viral RNA remained stable up to 168h, and there were no statistically significant differences among the needle samples. CONCLUSIONS: Although the infectivity of the viral material in the needles is unknown, the data indicate the need to re-evaluate the practices in cases of occupational accidents in which the source is not identified.

Estimativa dos parâmetros de compactação de solos usando técnicas de data mining

Dissertação de mestrado integrado em Engenharia Civil

Reconstruction of the regulatory network for Bacillus subtilis and reconciliation with gene expression data

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fmicb. 2016.00275