Showing papers by "Gilles Pagès published in 1995"

On the A.S. Convergence of the Kohonen Algorithm with a General Neighborhood Function

Abstract: Some existence and stability results for the equilibrium points of the one-dimensional Kohonen self-organizing neural network with two neighbors are extended to most nonincreasing neighborhood functions. All the functions mentioned in the neural literature are included. The assumption on the stimuli distribution is weakened, too. In the multidimensional setting, we derive from a general formula various stability and instability results.

B-Raf protein isoforms interact with and phosphorylate Mek-1 on serine residues 218 and 222.

Abstract: The B-raf/c-Rmil proto-oncogene belongs to the raf/mil family of serine/threonine protein kinases. It encodes multiple protein isoforms resulting from alternative splicing of two exons located upstream of the kinase domain. Recent studies suggested that B-Raf could be the intermediate molecule between Ras and Mek-1 (MAP Kinase Kinase) in signalling pathways specific of neural cells. However, there has been no evidence for a direct interaction between B-Raf and Mek-1. We report here that different B-Raf isoforms can be co-immunoprecipitated with anti-Mek-1 antisera in COS-1 cells and that the kinase activity of B-Raf is not required for its interaction with Mek-1. We also show that all B-Raf isoforms tested phosphorylate Mek-1 in a time-dependent manner, whereas kinase defective mutants fail to do so. Finally, we demonstrate that the constitutively activated S218D, S222D and S218D/S222D mutants of Mek-1 interact similarly with B-Raf. However, only the S218D and S222D mutants, and not the S218D/S222D double mutant, can be phosphorylated by B-Raf isoforms. Therefore, serine residues 218 and 222, previously shown to regulate Mek-1 activity, appear to be the major phosphorylation sites by B-Raf in vitro.

[MAP kinase module: role in the control of cell proliferation].

Abstract: Une cascade de proteines kinases tres conservee au cours de l'evolution, Raf/MAP kinase kinase kinase (MAPKKK) - MAP kinase kinase (MAPKK) → MAP kinase (MAPK) → ribosomal S6 kinase (p90 RSK), joue un role crucial dans la transmission des signaux externes de la membrane vers le noyau. Dans les cellules de mammiferes, cette cascade est activee a la fois par les recepteurs a activite tyrosine kinase et les recepteurs couples aux proteines G. Bien que le mode d'activation au niveau des recepteurs differe, tous les mitogenes activent les deux isoformes ubiquitaires de MAPK p42 et p44. Nous avons clone, etiquete et exprime dans les fibroblastes de Hamster la MAPKK et la MAPK de forme p44, afin d'etudier leur activation en fonction du temps, leur localisation subcellulaire, leur mode de regulation et leur role dans la proliferation cellulaire. L'activation des MAPK est rapide, biphasique et persistante, mais seuls les agents mitogenes puissants l'activent de maniere prolongee. L'activation de la MAPKK est egalement rapide et persistante mais ne fait pas apparaitre de difference entre les agents mitogenes et non mitogenes. Cette distinction s'etablit donc au niveau des MAP kinases, ce qui suggere l'intervention de phosphatases specifiques, comme la MAPK phosphatase MKP-1. Les deux isoformes de la MAPK, p42 et p44 sont relocalisees dans le noyau en reponse aux facteurs de croissance, contrairement aux elements en amont (MAPKK, MAPKKK et c-Raf) qui demeurent cytoplasmiques. Cette translocation nucleaire, ainsi que la capacite des MAPK a phosphoryler des facteurs de transcription, indiquent que la MAPK est un relais important entre les evenements cytoplasmiques et les evenements nucleaires. Enfin, le blocage de l'activation des MAPK obtenu, soit par expression de mutants inactifs ou d'un antisens de la MAPK p44, soit par surexpression de la MAPK phosphatase specifique MKP-1, inhibe la proliferation cellulaire. Parallelement, l'activation constitutive de la MAPKK reduit le besoin en facteurs de croissance, permet la croissance autonome et induit la formation de tumeurs chez les souris nude. L'activation des MAP kinases est donc a la fois necessaire et suffisante pour declencher l'entree dans le cycle cellulaire et la proliferation.

Comments about "Analysis of the convergence properties of topology preserving neural networks"

Abstract: Shows that the main proofs of the above paper (Yu et al., Trans Neural Networks, vol. 4, no. 2, p. 207-220, 1993) are incomplete and not correct: in fact, the self-organization cannot be achieved if the adaptation parameter satisfies the classical Robins-Monro conditions and Proposition 2 is erroneous. On the other hand, the two-dimensional extension (Theorem 3) is not proved. The main point is that the four classes that the authors consider as stable classes are not stable at all. Some references are finally given. >

About the Kohonen algorithm: strong or weak self-organization?

Abstract: The question of self-organization for the Kohonen algorithm is investigated. First the notions of organized states, weak and strong self-organizations are precisely defined. Then, combining mathematical and simulation results we prove that the Kohonen algorithm has not the strong self-organization property at least in two well-known cases: the stimuli space is [0, 1]2, the unit set is a line (resp. a grid) with the two nearest (resp. eight nearest) neighbourhood function. Copyright © 1996 Elsevier Science Ltd

Approximation of functions by perceptrons: a new approach

Abstract: We provide a constructive proof of the theorem of function approximation by perceptrons (cf Leshno et al. [1], Hornik [2]) when the activation function ψ isC∞ with all its derivatives non 0 at 0. We deal with uniform approximation on compact sets of continuous functions on ℜd,d≥1. This approach is elementary and provides some approximation results for the derivatives along with some bounds for the hidden layer.