A new Jacobian matrix for optimal learning of single-layer neural networks

This paper investigates the learning of a wide class of single-hidden-layer feedforward neural networks (SLFNs) with two sets of adjustable parameters, i.e., the nonlinear parameters in the hidden nodes and the linear output weights. The main objective is to both speed up the convergence of second-order learning algorithms such as Levenberg-Marquardt (LM), as well as to improve the network performance. This is achieved here by reducing the dimension of the solution space and by introducing a new Jacobian matrix. Unlike conventional supervised learning methods which optimize these two sets of parameters simultaneously, the linear output weights are first converted into dependent parameters, thereby removing the need for their explicit computation. Consequently, the neural network (NN) learning is performed over a solution space of reduced dimension. A new Jacobian matrix is then proposed for use with the popular second-order learning methods in order to achieve a more accurate approximation of the cost function. The efficacy of the proposed method is shown through an analysis of the computational complexity and by presenting simulation results from four different examples.

Development of microsatellites for the invasive riparian plant Impatiens glandulifera (Himalayan balsam) using intersimple sequence repeat cloning

Impatiens glandulifera (Himalayan balsam) is an invasive riparian plant species that can outcompete native perennials. Population genetic data on dispersal may aid in the management of invasive species, so we have developed microsatellite markers for this significant invader using an intersimple sequence repeat (ISSR)-based cloning method. Eight polymorphic markers displayed between two and five alleles, with overall levels of observed and expected heterozygosities ranging from 0.0500 to 0.7500 and from 0.1449 to 0.7692, respectively.

Development of microsatellites for the peat moss Sphagnum capillifolium using ISSR cloning

Sphagnum mosses are major components of peat bogs but populations of many species are under threat due to habitat fragmentation resulting from the cutting of peat for fuel. We have used an intersimple sequence repeat (ISSR)-based cloning method to develop nine polymorphic nuclear microsatellites for the peat moss species Sphagnum capillifolium. Between three and seven alleles per locus were detected in a sample of 48 haploid gametophytes and levels of gene diversity ranged from 0.5391 to 0.7960. These represent the first microsatellite markers developed for this important genus and most also exhibited cross-species amplification across a range of common Sphagnum species.

Kassina senegalensis skin tachykinins: Molecular cloning of kassinin and (Thr2, Ile9)-kassinin biosynthetic precursor cDNAs and comparative bioactivity of mature tachykinins on the smooth muscle of rat urinary bladder

Tachykinins are among the most widely-studied families of regulatory peptides characterized by a highly-conserved C-terminal -Phe-X-Gly-Leu-Met.amide motif, which also constitutes the essential bioactive core. The amphibian skin has proved to be a rich source of these peptides with physalaemin from the skin of Physalaemus fuscomaculatus representing the archetypal aromatic tachykinin (X = Tyr or Phe) and kassinin from the skin of Kassina senegalensis representing the archetypal aliphatic tachykinin in which X = Val or Ile. Despite the primary structures of both mature peptides having been known for at least 30 years, neither the structures nor organizations of their biosynthetic precursors have been reported. Here we report the structure and organization of the biosynthetic precursor of kassinin deduced from cDNA cloned from a skin secretion library. In addition, a second precursor cDNA encoding the novel kassinin analog (Thr2, Ile9)-kassinin was identified as was the predicted mature peptide in skin secretion. Both transcripts exhibited a high degree of nucleotide sequence similarity and of open-reading frame translated amino acid sequences of putative precursor proteins. The translated preprotachykinins each consisted of 80 amino acid residues encoding single copies of either kassinin or its site-substituted analog. Synthetic replicates of each kassinin were found to be active on rat urinary bladder smooth muscle at nanomolar concentrations. The structural organization of both preprotachykinins differs from that previously reported for those of Odorrana grahami skin indicating a spectrum of diversity akin to that established for amphibian skin preprobradykinins.

Cloning and characterization of a novel cis-naphthalene dihydrodiol dehydrogenase gene (narB) from Rhodococcus sp NCIMB12038

Rhodococcus sp. NCIMB112038 can utilize naphthalene as its sole carbon and energy source. The gene encoding cis-naphthalene dihydrodiol dehydrogenase (narB) of this strain has been cloned and sequenced. Expression of NCIMB12038 cis-naphthalene dihydrodiol dehydrogenase was demonstrated in Escherichia coli cells. narB encodes a putative protein of 271 amino acids and shares 39% amino acid identity with the cis-naphthalene dihydrodiol dehydrogenase from Pseudomonas putida G7. Comparison of NarB with some putative cis-dihydrodiol dehydrogenases from Rhodococcus species revealed significant differences between these proteins. NarB together with two other proteins forms a new group of cis-dihydrodiol dehydrogenases. (C) 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.