Cloning, sequencing and characterization of a chitin synthase gene fragment from the fungus Phascolomyces articulosus /

Phascolomyces articulosus genomic DNA was isolated from 48 h old hyphae and was used for amplification of a chitin synthase fragment by the polymerase chain reaction method. The primers used in the amplification corresponded to two widely conserved amino acid regions found in chitin synthases of many fimgi. Amphfication resulted in four bands (820, 900, 1000 and 1500 bp, approximately) as visualized in a 1.2% agarose gel. The lowest band (820 bp) was selected as a candidate for chitin synthase because most amplified regions from other fimgi so far exhibited similar sizes (600-750 bp). The selected fragment was extracted from the gel and cloned in the Hinc n site of pUC19. The derived plasmid and insert were designated ^\5C\9'PaCHS and PaCHS respectively. The plasmid pUC19-PaC/fS was digested by several restriction enzymes and was found to contain BamHl and HincU sites. Sequencing of PaCHS revealed two intron sequences and a total open reading frame of 200 amino acids. The derived polypeptide was compared with other related sequences from the EMBL database (Heidelberg, Germany) and was matched to 36 other fiilly or partially sequenced fimgal chitin synthase genes. The closest resemblance was with two genes (74.5% and 73.1% identity) from Rhizopus oligosporus. Southern hybridization with the cloned fragment as a probe to the PCR reaction showed a strong signal at the fragment selected for cloning and weaker signals at the other two fragments. Southern hybridization with partially digested Phascolomyces articulosus genomic DNA showed a single band. The amino acid sequence was compared with sequences from other chitin synthase gene classes using the CLUSTALW program. The chitin synthase fragment from Phascolomyces articulosus was initially grouped in class n along with chitin synthase fragments from Rhizopus oligosporus and Phycomyces blakesleeanus which also belong to the same class, Zygomycetes. Bootstrap analysis using the neighbor-joining method available by CLUSTALW verified such classification. Comparison of PaCHS revealed conservation of intron positions that are characteristic of chitin synthase gene fragments of zygomycetous fungi.

Construction of bovine adenovirus type 3 E1 and E3, substitution plasmids and the sequencing analysis of DNA pol and pTP /

The relative ease to concentrate and purify adenoviruses, their well characterized mid-sized genome, and the ability to delete non-essential regions from their genome to accommodate foreign gene, made adenoviruses a suitable candidate for the construction of vectors. The use of adenoviral vectors in gene therapy, vaccination, and as a general vector system for expressing foreign genes have been documented for some time. In this study, the objective was to rescue a BAV3 E1 or E3 recombinant vector carrying the kanamycin resistant gene, a dominant selectable marker with useful applications in studying vectored gene expression in mammalian cells. To accomplish the objective of this study, more information about BAV3 DNA sequences was required in order to make the manipulation of the virus genome accessible. Therefore, sequencing of the BAV3 genome from 1 1 .7% to 30.8% was carried out. Analysis of the determined sequences revealed the primary structure of important viral gene products coded by E2 including BAV3 DNA pol and precursor to terminal protein. Comparative analysis of these proteins with their counterparts from human and non human adenoviruses revealed important insights as to the evolutionary lineage of BAV3. In order to insert the kanamycin resistance gene in either E1 or E3, it was necessary to delete BAV3 sequences to accommodate the foreign gene so as not to exceed the limit of the packaging capacity of the virus. To construct a recombinant BAV3 in which a foreign gene was inserted in the deleted E1 region, an E1 shuttle vector was constructed. This involved the deletion from the viral sequences a region between 1.3% to 9% and inserting the kanamycin resistance gene to replace the deletion. The E1 shuttle vector contained the left (0%- 53.9%) segment of the genome and was expected to generate BAV3 recombinants that can be grown and propagated in cells that can complement the missing E1 functions. To construct a similar shuttle vector for E3 deletion, DNA sequences extending from 78.9% to 82.5% (1281 bp) were deleted from within the E3 region that had been cloned into a plasmid vector. The deleted region corresponds to those that have been shown to be non-essential for viral replication in cell culture. The resulting plasmid was used to construct another recombinant plasmid with BAV3 DNA sequences extending from 37.1% to 100% and with a deletion of E3 sequences that were replaced by kanamycin resistance gene. This shuttle plasmid was used in cotransfections with digested viral DNA in an attempt to rescue a recombinant BAV3 carrying the kanamycin resistance gene to replace the deleted E3. In spite of repeated attempts of transfection, El or E3 recombinant BAV3 were not isolated. It seems that other approaches should be applied to make a final conclusion on BAV3 infectivity.

Locating and sequencing of the E3 and neighbouring regions in bovine advenovirus type 2

Adenoviruses are nonenveloped icosahedral shaped particles. The double stranded DNA viral genome is divided into 5 major early transcription units, designated E1 A, E1 B, and E2 to E4, which are expressed in a regulated manner soon after infection. The gene products of the early region 3 (E3), shown to be nonessential for viral replication in vitro, are believed to be involved in counteracting host immunosurveillance. In order to sequence the E3 region of Bovine adenovirus type 2 (BAV2) it was necessary to determine the restriction map for the plasmid pEA48. A physical restriction endonuclease map for BamHl, Clal, Eco RI, Hindlll, Kpnl, Pstt, Sail, and Xbal was constructed. The DNA insert in pEA48 was determined to be viral in origin using Southern hybridization. A human adenovirus type 5 recombinant plasmid, containing partial DNA fragments of the two transcription units L4 and L5 that lie just outside the E3, was used to localize this region. The recombinant plasmid pEA was subcloned to facilitate sequencing. The DNA sequences between 74.8 and 90.5 map units containing the E3, the hexon associated protein (pVIII), and the fibre gene were determined. Homology comparison revealed that the genes for the hexon associated pV11I and the fibre protein are conserved. The last 70 amino acids of the BAV2 pV11I were the most conserved, showing a similarity of 87 percent with Ad2 pV1I1. A comparison between the predicted amino acid sequences of BAV2 and Ad40, Ad41 , Ad2 and AdS, revealed that they have an identical secondary structure consisting of a tail, a shaft and a knob. The shaft is composed of 22, 15 amino acid motifs, with periodic glycines and hydrophobic residues. The E3 region was found to consist of about 2.3 Kbp and to encode four proteins that were greater than 60 amino acids. However, these four open reading frames did not show significant homology to any other known adenovirus DNA or protein sequence.

The cloning and reconstitution of a bovine adenovirus type 2 E3 deletion mutant and the sequencing and analysis of the early 4 region

Recombinant Adenoviruses (Ads) have been shown to have potential applications in three areas: gene therapy, high level protein expression and recombinant vaccines.' At least three different locations within the Ad genome can be deleted and subsequently used for the insertion of foreign sequences. These include the Early 3 (E3), Early 1 (E1) and Early 4 (E4) regions. Viral vectors of this type have been well studied in Human Ads 2 and 5, however one has not yet been constructed for Bovine Adenovirus Type 2 (BAV2). The E3 region is located between 76.6 and 86 m.u. on the r-strand and is transcribed in a rightward direction. The gene products of the Early 3 region (E3) have been shown to be non-essential for viral replication, in vitro, but are required for host immunosurveillance. This study represents the cloning and reconstitution of a BAV2 E3 deletion mutant. A deletion of 1800bp was made within the E3 region of BAV2 and the thymidine kinase gene was subsequently inserted in the deleted area . . The plasmid pdlE3-4tk1 (23.4Kbp) was constructed and used to to facilitate homologous recombination with the wild type BAV2 to produce a mutant. Southern Blotting and Hybridization results suggest the presence of a BAV2 E3 deletion mutant with thymidine kinase sequences present. The E4 region of Human Adenovirus types 2 and 5 is located at the extreme right end of the genome (91.3 map units - 99.1 map units) and is transcribed in a leftward direction giving rise to a complicated set of differentially spliced mRNAs. Essentially there are 7 open reading frames (ORFs) encoding for at least 7 polypeptides. The gene products encoded by the E4 region have been shown to be essential for the expression of late viral genes, host cell shutoff and normal viral growth. We have cloned and sequenced the right end segment between 90.5 map units and 100 map units of the BAV2 genome. The results show several open reading frames which encode polypeptides exhibiting homology to three polypeptides encoded by the E4 region of human adenovirus type 2. These include the 14kDa protein encoded by ORF1, the 34kDa protein encoded by ORF6 and the 13kDa protein encoded by ORF3. The nucleotide sequence, restriction enzyme map, and ORF map of the E4 region could be very useful in future molecular manipulation of this region and could possibly explain the slow growth rate of BAV2 in MDBK cells.

The law of the sea: the United States, Canada and the deep seabed

The Falkland Islands War of 1982 was fought over competing claims to sovereignty over a group of islands off the east coast of South America. The dispute was between Argentina and the United Kingdom. Argentina claims the islands under rights to Spanish succession, the fact that they lie off the Argentine coast line and that in 1833 Great Britain took the islands illegally and by force. The United Kingdom claims the islands primarily through prescription--the fact that they have governed the islands in a peaceful, continuous and public manner since 1833. The British also hold that the population living on the islands, roughly eighteen hundred British descendants, should be able to decide their own future. The United Kingdom also lays claim to the islands through rights of discovery and settlement, although this claim has always been challenged by Spain who until 1811 governed the islands. Both claims have legal support, and the final decision if there will ever be one is difficult to predict. Sadly today the ultimate test of sovereignty does not come through international law but remains in the idea that "He is sovereign who can defend his sovereignty." The years preceding the Argentine invasion of 1982 witnessed many diplomatic exchanges between The United Kingdom and Argentina over the future of the islands. During this time the British sent signals to Argentina that ii implied a decline in British resolve to hold the islands and demonstrated that military action did more to further the talks along than did actual negotiations. The Argentine military junta read these signals and decided that they could take the islands in a quick military invasion and that the United Kingdom would consider the act as a fait accompli and would not protest the invasion. The British in response to this claimed that they never signaled to Argentina that a military solution was acceptable to them and launched a Royal Navy task force to liberate the islands. Both governments responded to an international crisis with means that were designed both to resolve the international crisis and increase the domestic popularity of the government. British Prime Minister Margaret Thatcher was facing an all-time low in popularity for post-War Prime Ministers while Argentine President General Galtieri needed to gain mass popular support so he could remain a viable President after he was scheduled to lose command of the army and a seat on the military junta that ran the country. The military war for the Falklands is indicative of the nature of modern warfare between Third World countries. It shows that the gap in military capabilities between Third and First World countries is narrowing significantly. Modern warfare between a First and Third World country is no longer a 'walk over' for the First World country.

The wars of the gulls; : an historical romance in three chapters; chap. I, Shewing how and why and with whom the gulls went to war: chap. II, Shewing how the gulls make the deep to boil like a pot: chap. III, Shewing how a certain doughty general of the gulls goes forth to play the game of hull-gull in Upper Canada.

Transcribed on front paste-down: W.G. Phelps Oct. 29 1890.

Second Welland Canal - Book 2, Survey Map 8 - Little Deep Cut in Thorold

Survey map of the Second Welland Canal created by the Welland Canal Company showing the canal in Thorold South. Identified structures associated with the Canal include the Little Deep Cut and the towing path. The surveyors' measurements and notes can be seen in red and black ink and pencil. Local area landmarks are also identified and include streets and roads (ex. Road to Beaverdams and Road to Allanburgh), two unnamed bridges, the Spoil Bank, a pond, and the Back Water. Properties and property owners of note are: Lots 29 and 30, Jacob Keefer, John Brown, William Bouck, C. Gisso, and a property reserved for Bridge Tender.

Second Welland Canal - Book 2, Survey Map 15 - Deep Cut in Thorold

Survey map of the Second Welland Canal created by the Welland Canal Company showing the canal in the Thorold Township just south of Allanburgh. Identified structures and features associated with the Canal include the Deep Cut and the towing path. The surveyors' measurements and notes can be seen in red and black ink and pencil. Local area landmarks are also identified and include streets and roads (ex. Road to Port Robinson), and the Spoil Bank. Properties and property owners of note are: Lots 142 and 143, John J. Church, Henry Vanderburgh, and Martin Delamatter and G. Coulter.

Second Welland Canal - Book 2, Survey Map 16 - Deep Cut in Thorold

Survey map of the Second Welland Canal created by the Welland Canal Company showing the canal in the Thorold Township between Allanburg and Port Robinson. Identified structures and features associated with the Canal include the Deep Cut and the towing path. The surveyors' measurements and notes can be seen in red and black ink and pencil. Local area landmarks are also identified and include streets and roads (ex. Road to Port Allanburg), and the Spoil Bank. Properties and property owners of note are: Lots 185, 186, and 187, J. J. Church and H. Vanderburgh. Four properties adjacent to the canal are outlined in blue and labeled J through M, with L and K belonging to John Beatty, M belonging to John Coulter, and J belonging to G. Jordan (formerly belonging to John Coleman Jordan).

Second Welland Canal - Book 2, Survey Map 17 - Deep Cut and Port Robinson

Survey map of the Second Welland Canal created by the Welland Canal Company showing the canal at Port Robinson. Identified structures and features associated with the Canal include the Deep Cut, Old Channel of Canal, and the towing path. The surveyors' measurements and notes can be seen in red and black ink and pencil. Local area landmarks are also identified and include streets and roads (ex. Road to Port Allanburg), the Spoil Bank, an island, several bridges, and a church. Several unidentified structures are present but not labeled. Properties and property owners of note are: Lots 202, 203, and 204. Lot 203 is divided into several properties labeled A - J. Owners of these properties include James McCoppen, John Coulter, James Griffith, John C. Jordan, W. Hendershot, John Greer, Charles Richards, C. Stuart, and S. D. Woodruff. Other property owners include D. McFarland.

Chart of approximate quantity of excavation in slides in the deep cut

Chart of approximate quantity of excavation in slides in the deep cut, July 1, 1848.

Training deep convolutional architectures for vision

Les tâches de vision artificielle telles que la reconnaissance d’objets demeurent irrésolues à ce jour. Les algorithmes d’apprentissage tels que les Réseaux de Neurones Artificiels (RNA), représentent une approche prometteuse permettant d’apprendre des caractéristiques utiles pour ces tâches. Ce processus d’optimisation est néanmoins difficile. Les réseaux profonds à base de Machine de Boltzmann Restreintes (RBM) ont récemment été proposés afin de guider l’extraction de représentations intermédiaires, grâce à un algorithme d’apprentissage non-supervisé. Ce mémoire présente, par l’entremise de trois articles, des contributions à ce domaine de recherche. Le premier article traite de la RBM convolutionelle. L’usage de champs réceptifs locaux ainsi que le regroupement d’unités cachées en couches partageant les même paramètres, réduit considérablement le nombre de paramètres à apprendre et engendre des détecteurs de caractéristiques locaux et équivariant aux translations. Ceci mène à des modèles ayant une meilleure vraisemblance, comparativement aux RBMs entraînées sur des segments d’images. Le deuxième article est motivé par des découvertes récentes en neurosciences. Il analyse l’impact d’unités quadratiques sur des tâches de classification visuelles, ainsi que celui d’une nouvelle fonction d’activation. Nous observons que les RNAs à base d’unités quadratiques utilisant la fonction softsign, donnent de meilleures performances de généralisation. Le dernière article quand à lui, offre une vision critique des algorithmes populaires d’entraînement de RBMs. Nous montrons que l’algorithme de Divergence Contrastive (CD) et la CD Persistente ne sont pas robustes : tous deux nécessitent une surface d’énergie relativement plate afin que leur chaîne négative puisse mixer. La PCD à "poids rapides" contourne ce problème en perturbant légèrement le modèle, cependant, ceci génère des échantillons bruités. L’usage de chaînes tempérées dans la phase négative est une façon robuste d’adresser ces problèmes et mène à de meilleurs modèles génératifs.

Understanding deep architectures and the effect of unsupervised pre-training

Cette thèse porte sur une classe d'algorithmes d'apprentissage appelés architectures profondes. Il existe des résultats qui indiquent que les représentations peu profondes et locales ne sont pas suffisantes pour la modélisation des fonctions comportant plusieurs facteurs de variation. Nous sommes particulièrement intéressés par ce genre de données car nous espérons qu'un agent intelligent sera en mesure d'apprendre à les modéliser automatiquement; l'hypothèse est que les architectures profondes sont mieux adaptées pour les modéliser. Les travaux de Hinton (2006) furent une véritable percée, car l'idée d'utiliser un algorithme d'apprentissage non-supervisé, les machines de Boltzmann restreintes, pour l'initialisation des poids d'un réseau de neurones supervisé a été cruciale pour entraîner l'architecture profonde la plus populaire, soit les réseaux de neurones artificiels avec des poids totalement connectés. Cette idée a été reprise et reproduite avec succès dans plusieurs contextes et avec une variété de modèles. Dans le cadre de cette thèse, nous considérons les architectures profondes comme des biais inductifs. Ces biais sont représentés non seulement par les modèles eux-mêmes, mais aussi par les méthodes d'entraînement qui sont souvent utilisés en conjonction avec ceux-ci. Nous désirons définir les raisons pour lesquelles cette classe de fonctions généralise bien, les situations auxquelles ces fonctions pourront être appliquées, ainsi que les descriptions qualitatives de telles fonctions. L'objectif de cette thèse est d'obtenir une meilleure compréhension du succès des architectures profondes. Dans le premier article, nous testons la concordance entre nos intuitions---que les réseaux profonds sont nécessaires pour mieux apprendre avec des données comportant plusieurs facteurs de variation---et les résultats empiriques. Le second article est une étude approfondie de la question: pourquoi l'apprentissage non-supervisé aide à mieux généraliser dans un réseau profond? Nous explorons et évaluons plusieurs hypothèses tentant d'élucider le fonctionnement de ces modèles. Finalement, le troisième article cherche à définir de façon qualitative les fonctions modélisées par un réseau profond. Ces visualisations facilitent l'interprétation des représentations et invariances modélisées par une architecture profonde.

On Recurrent and Deep Neural Networks

L'apprentissage profond est un domaine de recherche en forte croissance en apprentissage automatique qui est parvenu à des résultats impressionnants dans différentes tâches allant de la classification d'images à la parole, en passant par la modélisation du langage. Les réseaux de neurones récurrents, une sous-classe d'architecture profonde, s'avèrent particulièrement prometteurs. Les réseaux récurrents peuvent capter la structure temporelle dans les données. Ils ont potentiellement la capacité d'apprendre des corrélations entre des événements éloignés dans le temps et d'emmagasiner indéfiniment des informations dans leur mémoire interne. Dans ce travail, nous tentons d'abord de comprendre pourquoi la profondeur est utile. Similairement à d'autres travaux de la littérature, nos résultats démontrent que les modèles profonds peuvent être plus efficaces pour représenter certaines familles de fonctions comparativement aux modèles peu profonds. Contrairement à ces travaux, nous effectuons notre analyse théorique sur des réseaux profonds acycliques munis de fonctions d'activation linéaires par parties, puisque ce type de modèle est actuellement l'état de l'art dans différentes tâches de classification. La deuxième partie de cette thèse porte sur le processus d'apprentissage. Nous analysons quelques techniques d'optimisation proposées récemment, telles l'optimisation Hessian free, la descente de gradient naturel et la descente des sous-espaces de Krylov. Nous proposons le cadre théorique des méthodes à région de confiance généralisées et nous montrons que plusieurs de ces algorithmes développés récemment peuvent être vus dans cette perspective. Nous argumentons que certains membres de cette famille d'approches peuvent être mieux adaptés que d'autres à l'optimisation non convexe. La dernière partie de ce document se concentre sur les réseaux de neurones récurrents. Nous étudions d'abord le concept de mémoire et tentons de répondre aux questions suivantes: Les réseaux récurrents peuvent-ils démontrer une mémoire sans limite? Ce comportement peut-il être appris? Nous montrons que cela est possible si des indices sont fournis durant l'apprentissage. Ensuite, nous explorons deux problèmes spécifiques à l'entraînement des réseaux récurrents, à savoir la dissipation et l'explosion du gradient. Notre analyse se termine par une solution au problème d'explosion du gradient qui implique de borner la norme du gradient. Nous proposons également un terme de régularisation conçu spécifiquement pour réduire le problème de dissipation du gradient. Sur un ensemble de données synthétique, nous montrons empiriquement que ces mécanismes peuvent permettre aux réseaux récurrents d'apprendre de façon autonome à mémoriser des informations pour une période de temps indéfinie. Finalement, nous explorons la notion de profondeur dans les réseaux de neurones récurrents. Comparativement aux réseaux acycliques, la définition de profondeur dans les réseaux récurrents est souvent ambiguë. Nous proposons différentes façons d'ajouter de la profondeur dans les réseaux récurrents et nous évaluons empiriquement ces propositions.