Artificial neural network models for prediction of daily fine particulate matter concentrations in Algiers.

Neural network (NN) models were evaluated for the prediction of suspended particulates with aerodynamic diameter less than 10-µm (PM10) concentrations. The model evaluation work considered the sequential hourly concentration time series of PM10, which were measured at El Hamma station in Algiers. Artificial neural network models were developed using a combination of meteorological and time-scale as input variables. The results were rather satisfactory, with values of the coefficient of correlation (R (2)) for independent test sets ranging between 0.60 and 0.85 and values of the index of agreement (IA) between 0.87 and 0.96. In addition, the root mean square error (RMSE), the mean absolute error (MAE), the normalized mean squared error (NMSE), the absolute relative percentage error (ARPE), the fractional bias (FB), and the fractional variance (FS) were calculated to assess the performance of the model. It was seen that the overall performance of model 3 was better than models 1 and 2.

Successful treatment of intra-abdominal eumycotic mycetoma caused by Penicillium duponti in a dog.

A 2-year-old female neutered golden retriever was presented for investigation of an intra-abdominal mass. Computed tomography revealed a mass associated with the caudal pole of the right kidney. Incisional biopsy findings were consistent with eumycotic mycetoma. The mass was subsequently removed in conjunction with right ureteronephrectomy. Two years later, the dog re-presented with a splenic mass and fungal plaques located throughout the peritoneum. Splenectomy was performed and the mass was diagnosed as eumycotic mycetoma caused by Penicillium duponti. Indefinite systemic treatment with 10 mg/kg itraconazole orally once a day was initiated. Thirty-two months after the last surgery, there were no clinical signs apart from mild polydipsia. Haematology and biochemistry results were unremarkable. To the authors' knowledge, this is the first report of successful treatment of intra-abdominal eumycotic mycetoma with a combination of surgery and systemic antifungal therapy in the dog. Penicillium duponti has not apparently been reported to cause disease in animals or humans.

Testing candidate plant barcode regions in the Myristicaceae.

The concept and practice of DNA barcoding have been designed as a system to facilitate species identification and recognition. The primary challenge for barcoding plants has been to identify a suitable region on which to focus the effort. The slow relative nucleotide substitution rates of plant mitochondria and the technical issues with the use of nuclear regions have focused attention on several proposed regions in the plastid genome. One of the challenges for barcoding is to discriminate closely related or recently evolved species. The Myristicaceae, or nutmeg family, is an older group within the angiosperms that contains some recently evolved species providing a challenging test for barcoding plants. The goal of this study is to determine the relative utility of six coding (Universal Plastid Amplicon - UPA, rpoB, rpoc1, accD, rbcL, matK) and one noncoding (trnH-psbA) chloroplast loci for barcoding in the genus Compsoneura using both single region and multiregion approaches. Five of the regions we tested were predominantly invariant across species (UPA, rpoB, rpoC1, accD, rbcL). Two of the regions (matK and trnH-psbA) had significant variation and show promise for barcoding in nutmegs. We demonstrate that a two-gene approach utilizing a moderately variable region (matK) and a more variable region (trnH-psbA) provides resolution among all the Compsonuera species we sampled including the recently evolved C. sprucei and C. mexicana. Our classification analyses based on nonmetric multidimensional scaling ordination, suggest that the use of two regions results in a decreased range of intraspecific variation relative to the distribution of interspecific divergence with 95% of the samples correctly identified in a sequence identification analysis.