Generic optimality conditions for semi-algebraic convex programs

We consider linear optimization over a nonempty convex semi-algebraic feasible region F. Semidefinite programming is an example. If F is compact, then for almost every linear objective there is a unique optimal solution, lying on a unique \active" manifold, around which F is \partly smooth", and the second-order sufficient conditions hold. Perturbing the objective results in smooth variation of the optimal solution. The active manifold consists, locally, of these perturbed optimal solutions; it is independent of the representation of F, and is eventually identified by a variety of iterative algorithms such as proximal and projected gradient schemes. These results extend to unbounded sets F.

A critical role for the T cell receptor alpha-chain connecting peptide domain in positive selection of CD1-independent NKT cells.

Natural killer T (NKT) cells are a subset of mature alpha beta TCR(+) cells that co-express NK lineage markers. Whereas most NKT cells express a canonical Valpha14/Vbeta8.2 TCR and are selected by CD1d, a minority of NKT cells express a diverse TCR repertoire and develop independently of CD1d. Little is known about the selection requirements of CD1d-independent NKT cells. We show here that NKT cells develop in RAG-deficient mice expressing an MHC class II-restricted transgenic TCR (Valpha2/Vbeta8.1) but only under conditions that lead to negative selection of conventional T cells. Moreover development of NKT cells in these mice is absolutely dependent upon an intact TCR alpha-chain connecting peptide domain, which is required for positive selection of conventional T cells via recruitment of the ERK signaling pathway. Collectively our data demonstrate that NKT cells can develop as a result of high avidity TCR/MHC class II interactions and suggest that common signaling pathways are involved in the positive selection of CD1d-independent NKT cells and conventional T cells.

Molecular determinants for membrane association of the hepatitis C virus NS2 protease domain

Background: Hepatitis C virus (HCV) nonstructural protein 2 (NS2) plays essential roles in particle assembly and polyprotein processing. It harbors an N-terminal membrane domain comprising three putative transmembrane s egments ( amino acids [aa] 1-93) a nd a C-terminal cysteine protease domain (aa 94-217). Given that the latter has been predicted to be membrane-associated, we aimed to identify molecular determinants for membrane association of the NS2 protease domain. Methods: A comprehensive panel of NS2 deletion constructs was analyzed by fluorescence microscopy, selective membrane extraction, and m embrane flotation assays. Candidate aa r esidues involved in membrane association were substituted by site-directed mutagenesis. Results: The NS2 protease domain alone was found to associate with membranes. Two N-terminal α-helices comprising aa 102-114 and aa 123-136 were found to m ediate this a ssociation, w ith c onserved hydrophobic and positively charged aa residues representing the key determinants. I nterestingly, m utagenesis analyses r evealed that electrostatic interactions involving a positively charged aa residue in α-helix aa 123-136 are required for membrane association. Mono- and bicistronic (i.e. NS2 c leavage-independent) HCV constructs were prepared to i nvestigate the effect o f these substitutions on RNA replication and infectious viral particle formation. Conclusions: T he NS2 protease d omain itself harbors m olecular determinants for membrane association within α-helices aa 102-114 and aa 1 23-136 which may contribute to p roper p ositioning of t he active site. These results provide new insights i nto the membrane topology and t he p oorly understood f unction of t his essential viral protease.

Preferential binding of the methyl-CpG binding domain protein 2 at methylated transcriptional start site regions.

Methyl-CpG Binding Domain (MBD) proteins are thought to be key molecules in the interpretation of DNA methylation signals leading to gene silencing through recruitment of chromatin remodeling complexes. In cancer, the MBD-family member, MBD2, may be primarily involved in the repression of genes exhibiting methylated CpG at their 5' end. Here we ask whether MBD2 randomly associates methylated sequences, producing chance effects on transcription, or exhibits a more specific recognition of some methylated regions. Using chromatin and DNA immunoprecipitation, we analyzed MBD2 and RNA polymerase II deposition and DNA methylation in HeLa cells on arrays representing 25,500 promoter regions. This first whole-genome mapping revealed the preferential localization of MBD2 near transcription start sites (TSSs), within the region analyzed, 7.5 kb upstream through 2.45 kb downstream of 5' transcription start sites. Probe by probe analysis correlated MBD2 deposition and DNA methylation. Motif analysis did not reveal specific sequence motifs; however, CCG and CGC sequences seem to be overrepresented. Nonrandom association (multiple correspondence analysis, p < 0.0001) between silent genes, DNA methylation and MBD2 binding was observed. The association between MBD2 binding and transcriptional repression weakened as the distance between binding site and TSS increased, suggesting that MBD2 represses transcriptional initiation. This hypothesis may represent a functional explanation for the preferential binding of MBD2 at methyl-CpG in TSS regions.

Semi-supervised segmentation of ultrasound images based on patch representation and continuous min cut.

Ultrasound segmentation is a challenging problem due to the inherent speckle and some artifacts like shadows, attenuation and signal dropout. Existing methods need to include strong priors like shape priors or analytical intensity models to succeed in the segmentation. However, such priors tend to limit these methods to a specific target or imaging settings, and they are not always applicable to pathological cases. This work introduces a semi-supervised segmentation framework for ultrasound imaging that alleviates the limitation of fully automatic segmentation, that is, it is applicable to any kind of target and imaging settings. Our methodology uses a graph of image patches to represent the ultrasound image and user-assisted initialization with labels, which acts as soft priors. The segmentation problem is formulated as a continuous minimum cut problem and solved with an efficient optimization algorithm. We validate our segmentation framework on clinical ultrasound imaging (prostate, fetus, and tumors of the liver and eye). We obtain high similarity agreement with the ground truth provided by medical expert delineations in all applications (94% DICE values in average) and the proposed algorithm performs favorably with the literature.

Mutations at a single codon in Mad homology 2 domain of SMAD4 cause Myhre syndrome.

Myhre syndrome (MIM 139210) is a developmental disorder characterized by short stature, short hands and feet, facial dysmorphism, muscular hypertrophy, deafness and cognitive delay. Using exome sequencing of individuals with Myhre syndrome, we identified SMAD4 as a candidate gene that contributes to this syndrome on the basis of its pivotal role in the bone morphogenetic pathway (BMP) and transforming growth factor (TGF)-β signaling. We identified three distinct heterozygous missense SMAD4 mutations affecting the codon for Ile500 in 11 individuals with Myhre syndrome. All three mutations are located in the region of SMAD4 encoding the Mad homology 2 (MH2) domain near the site of monoubiquitination at Lys519, and we found a defect in SMAD4 ubiquitination in fibroblasts from affected individuals. We also observed decreased expression of downstream TGF-β target genes, supporting the idea of impaired TGF-β-mediated transcriptional control in individuals with Myhre syndrome.

Transcriptional Activation and Receptor Dimerization is Affected by Mutations in the NR2E3 Ligand-Binding Domain

Purpose: Mutations in the ligand-binding domain (LBD) of NR2E3 cause recessively inherited enhanced short wavelength sensitive (S-) cone syndrome (ESCS), Goldmann-Favre syndrome (GFS) and clumped pigmentary retinal degeneration (CPRD). In addition to ligand binding, the LBD contains also essential amino acid sequences for the oligomerization of nuclear receptors. The aim of our studies is to characterize the impact of mutations in the LBD on receptor oligomerization and transcriptional activity of NR2E3. Methods: The different NR2E3 mutants were generated by QuickChange mutagenesis and analyzed in 293T-based transactivation studies and BRET2 (bioluminescence resonance electron transfer) assays. In silico homology modeling of mutant proteins was also performed using available crystallographic data of related nuclear receptors. Results: The mutants p.W234S, p.A256V, p.A256E, p.L263P, p.R309G, p.R311Q, p.R334G, p.L336P, p.L353V, p.R385P and p.M407K, all located in the LBD, showed impaired receptor dimerization at various degrees. Impaired repressor dimerization as assessed by BRET2 assays did not always correlate with impaired repressor function of NR2E3 as assessed by cell-based reporter assays. There were minor differences of transcriptional activity of mutant proteins on mouse S-opsin (opn1sw), mouse cone arrestin (arr3) and human cone arrestin, suggesting that the effect of LBD mutations was independent of the promoter context. Conclusions: Mutational analysis and homology modeling allowed the characterization of potential oligomerization interfaces of the NR2E3 LBD. Additionally, mutations in NR2E3 LBD may cause recessive retinal degenerations by different molecular mechanisms.

Discrete time Non-homogeneous Semi-Markov Processes applied to Models for Disability Insurance

In this paper, we present a stochastic model for disability insurance contracts. The model is based on a discrete time non-homogeneous semi-Markov process (DTNHSMP) to which the backward recurrence time process is introduced. This permits a more exhaustive study of disability evolution and a more efficient approach to the duration problem. The use of semi-Markov reward processes facilitates the possibility of deriving equations of the prospective and retrospective mathematical reserves. The model is applied to a sample of contracts drawn at random from a mutual insurance company.

Regulation of the DNA-binding and transcriptional activities of Xenopus laevis NFI-X by a novel C-terminal domain.

The nuclear factor I (NFI) family consists of sequence-specific DNA-binding proteins that activate both transcription and adenovirus DNA replication. We have characterized three new members of the NFI family that belong to the Xenopus laevis NFI-X subtype and differ in their C-termini. We show that these polypeptides can activate transcription in HeLa and Drosophila Schneider line 2 cells, using an activation domain that is subdivided into adjacent variable and subtype-specific domains each having independent activation properties in chimeric proteins. Together, these two domains constitute the full NFI-X transactivation potential. In addition, we find that the X. laevis NFI-X proteins are capable of activating adenovirus DNA replication through their conserved N-terminal DNA-binding domains. Surprisingly, their in vitro DNA-binding activities are specifically inhibited by a novel repressor domain contained within the C-terminal part, while the dimerization and replication functions per se are not affected. However, inhibition of DNA-binding activity in vitro is relieved within the cell, as transcriptional activation occurs irrespective of the presence of the repressor domain. Moreover, the region comprising the repressor domain participates in transactivation. Mechanisms that may allow the relief of DNA-binding inhibition in vivo and trigger transcriptional activation are discussed.

MyHits: a new interactive resource for protein annotation and domain identification.

The MyHits web server (http://myhits.isb-sib.ch) is a new integrated service dedicated to the annotation of protein sequences and to the analysis of their domains and signatures. Guest users can use the system anonymously, with full access to (i) standard bioinformatics programs (e.g. PSI-BLAST, ClustalW, T-Coffee, Jalview); (ii) a large number of protein sequence databases, including standard (Swiss-Prot, TrEMBL) and locally developed databases (splice variants); (iii) databases of protein motifs (Prosite, Interpro); (iv) a precomputed list of matches ('hits') between the sequence and motif databases. All databases are updated on a weekly basis and the hit list is kept up to date incrementally. The MyHits server also includes a new collection of tools to generate graphical representations of pairwise and multiple sequence alignments including their annotated features. Free registration enables users to upload their own sequences and motifs to private databases. These are then made available through the same web interface and the same set of analytical tools. Registered users can manage their own sequences and annotations using only web tools and freeze their data in their private database for publication purposes.