Joint Optimization of Deep Neural Network-Based Dereverberation and Beamforming for Sound Event Detection in Multi-Channel Environments.

In this paper, we propose joint optimization of deep neural network (DNN)-supported dereverberation and beamforming for the convolutional recurrent neural network (CRNN)-based sound event detection (SED) in multi-channel environments. First, the short-time Fourier transform (STFT) coefficients are calculated from multi-channel audio signals under the noisy and reverberant environments, which are then enhanced by the DNN-supported weighted prediction error (WPE) dereverberation with the estimated masks. Next, the STFT coefficients of the dereverberated multi-channel audio signals are conveyed to the DNN-supported minimum variance distortionless response (MVDR) beamformer in which DNN-supported MVDR beamforming is carried out with the source and noise masks estimated by the DNN. As a result, the single-channel enhanced STFT coefficients are shown at the output and tossed to the CRNN-based SED system, and then, the three modules are jointly trained by the single loss function designed for SED. Furthermore, to ease the difficulty of training a deep learning model for SED caused by the imbalance in the amount of data for each class, the focal loss is used as a loss function. Experimental results show that joint training of DNN-supported dereverberation and beamforming with the SED model under the supervision of focal loss significantly improves the performance under the noisy and reverberant environments.

Detection of antibodies against SARS-Coronavirus using recombinant truncated nucleocapsid proteins by ELISA.

Severe acute respiratory syndrome (SARS) is a lifethreatening emerging respiratory disease caused by the coronavirus, SARS-CoV. The nucleocapsid (N) protein of SARS-CoV is highly antigenic and may be a suitable candidate for diagnostic applications. We constructed truncated recombinant N proteins (N1 [1-422 aa], N2 [1- 109 aa], and N3 [110-422 aa]) and determined their antigenicity by Western blotting using convalescent SARS serum. The recombinants containing N1 and N3 reacted with convalescent SARS serum in Western blotting. However, the recombinant with N2 did not. In ELISA using N1 or N3 as the antigens, positive results were observed in 10 of 10 (100%) SARS-CoV-positive human sera. None of 50 healthy sera gave positive results in either assay. These data indicate that the ELISA using N1 or N3 has high sensitivity and specificity. These results suggest that the middle or C-terminal region of the SARS N protein is important for eliciting antibodies against SARS-CoV during the immune response, and ELISA reactions using N1 or N3 may be a valuable tool for SARS diagnosis.

Quantitative GFP fluorescence as an indicator of arsenite developmental toxicity in mosaic heat shock protein 70 transgenic zebrafish.

In transgenic zebrafish (Danio rerio), green fluorescent protein (GFP) is a promising marker for environmental pollutants. In using GFP, one of the obstacles which we faced was how to compare toxicity among different toxicants or among a specific toxicant in different model species with the intensity of GFP expression. Using a fluorescence detection method, we first validated our method for estimating the amount of GFP fluorescence present in transgenic fish, which we used as an indicator of developmental toxicity caused by the well-known toxicant, arsenite. To this end, we developed mosaic transgenic zebrafish with the human heat shock response element (HSE) fused to the enhanced GFP (EGFP) reporter gene to indicate exposure to arsenite. We confirmed that EGFP expression sites correlate with gross morphological disruption caused by arsenite exposure. Arsenite (300.0 microM) caused stronger EGFP fluorescence intensity and quantity than 50.0 microM and 10.0 microM arsenite in our transgenic zebrafish. Furthermore, arsenite-induced apoptosis was demonstrated by TUNEL assay. Apoptosis was inhibited by the antioxidant, N-acetyl-cystein (NAC) in this transgenic zebrafish. The distribution of TUNEL-positive cells in embryonic tissues was correlated with the sites of arsenite toxicity and EGFP expression. The EGFP values quantified using the standard curve equation from the known GFP quantity were consistent with the arsenite-induced EGFP expression pattern and arsenite concentration, indicating that this technique can be a reliable and applicable measurement. In conclusion, we propose that fluorescence-based EGFP quantification in transgenic fish containing the hsp70 promoter-EGFP reporter-gene construct is a useful indicator of development toxicity caused by arsenite.

Arsenite-induced apoptosis is prevented by antioxidants in zebrafish liver cell line.

This study evaluated oxidative stress-induced apoptosis as a possible mechanism of arsenite toxicity in zebrafish liver cell line (ZFL cells). The heat shock protein 70 (HSP70), a chaperone protein, appears to provide protection against oxidative stress and apoptosis. Using the MTT assay, we demonstrated that survival of ZFL cells treated with arsenite for 24h decreased in a dose-dependent manner. The possible mechanisms that promote the cytotoxicity of arsenite were addressed. Cell viability assays revealed that arsenite caused a dose-dependent increase in cell death, and pretreatment of the ZFL cells with antioxidants blunted these effects. Antioxidants such as N-acetyl-cysteine (NAC, 5 mM) and dithiothreitol (DTT, 80 microM) significantly prevented ZFL cells from arsenite-induced death. Nuclear staining was performed using 1 microg/ml Hoechst, and cells were analyzed with a fluorescent microscope. Arsenite (30 microM) induced massive apoptosis that was identified by morphology and condensation and fragmentation of the nuclei of the ZFL cells. Pretreatment with NAC or DTT before arsenite insult effectively protected the cells against oxidative stress-induced apoptosis from the arsenite. Using a transfected human hsp 70 promoter-enhanced green fluorescent protein (EGFP) reporter, pHhsp70-EGFP, the induction of HSP70 against oxidative stress-induced apoptosis by arsenite was observed. The induction of HSP70 by arsenite increased in a dose-dependent manner, and pretreatment of transfected ZFL cells with NAC or DTT before arsenite insult reduced EGFP expression. Taken together, our results provide evidence that stimulation of the heat shock response is a sensitive biomarker of arsenic exposure and that arsenite causes oxidative stress-induced apoptosis in ZFL cells.