1,720,986 research outputs found

    Sparse logistic regression: Comparison of regularization and Bayesian implementations

    Full text link
    In knowledge-based systems, besides obtaining good output prediction accuracy, it is crucial to understand the subset of input variables that have most influence on the output, with the goal of gaining deeper insight into the underlying process. These requirements call for logistic model estimation techniques that provide a sparse solution, i.e., where coefficients associated with non-important variables are set to zero. In this work we compare the performance of two methods: the first one is based on the well known Least Absolute Shrinkage and Selection Operator (LASSO) which involves regularization with an l1 norm; the second one is the Relevance Vector Machine (RVM) which is based on a Bayesian implementation of the linear logistic model. The two methods are extensively compared in this paper, on real and simulated datasets. Results show that, in general, the two approaches are comparable in terms of prediction performance. RVM outperforms the LASSO both in term of structure recovery (estimation of the correct non-zero model coefficients) and prediction accuracy when the dimensionality of the data tends to increase. However, LASSO shows comparable performance to RVM when the dimensionality of the data is much higher than number of samples that is p >> n

    Etidronate inhibits osteoclast adhesion to bone surfaces but does not interfere with their specific recognition of single bone proteins

    No full text
    Bisphosphonates are nonbiodegradable pyrophosphate analogues that are increasingly used for inhibiting bone resorption in disorders characterized by excessive bone loss; they are in fact generally considered potent inhibitors of osteoclastic bone resorption both in vivo and in vitro but the mechanism by which these compounds exert their effect remains to be clarified. Data obtained on human systems in vitro are in particular lacking. The aim of this study was to assess the effect of etidronate (EHDP), one of the early compounds of bisphosphonates family, on osteoclast-matrix interactions and on bone resorption in a human in vitro system, using as an experimental model, osteoclast-like cells derived from giant cell tumors of bone (GCT). We reported that the presence of EHDP at the concentration ranging from 10-3 M to 10-7 M exerted a dose-dependent inhibition of the bone resorption activity with a maximal effect at 10-5 M. Because bisphosphonates owe the specificy of their actions to their ability of binding to bone surfaces, we performed adhesion assay using bone slices that had been pre-treated with solution of EHDP, at the established inhibitory concentration of hone resorption. Results showed that the morphology of cells plated onto bone slices pre-treated with the bisphosphonate was not significantly different from the control while the number of adherent cells was significantly reduced, by the treatment of about 50% vs control. Since osteoclast adhesion to the bone surface is mediated by the interaction with some adhesive proteins of extracellular bone matrix such as bone sialoprotein, osteopontin and fibronectin, furthermore the effect of EHDP on osteoclast adhesion onto specific extracellular matrix proteins, was also tested. In this case the presence of EHDP in the medium did not modify the percentage of cell adhesion compared to the control, indicating that the inhibitory effect of EHDP on cell adhesion onto hone slices, was probably not due to the interference with adhesion process of cells with specific hone matrix proteins. However we cannot exclude a possible effect on this process of EHDP that could be evident when all the components of the extracellular matrix ar present. In summary this work provides the first evidences of the inhibitory effect of EHDP on human osteoclast-like cells in vitro, confirming animal model data, possibly by a mechanism involving the adhesion process
    corecore