Search

It is a challenging task for computers to recognize humans’ emotions through their conversations. Therefore, this research is aimed at analyzing conversation audio data, then labeling humans’ emotions, finally annotating and visualizing the identified emotional traits of audio conversions in real-time. In order to make computer to process speech emotion features, the raw audio is converted from time domain to frequency domain and extract speech emotion by Mel-Frequency Cepstral Coefficients. In terms of speech emotion recognition, deep neural network and extreme learning machine are used to predict emotion traits. Each emotional trait is captured by speech recognition precision. There are four emotional traits which include sadness, happiness, neutral, anger in the dataset. The total precision value of four emotional traits is normalized into 1. In this study, the normalized precision is used as emotional trait relative intensity in which each emotional trait is labeled and displayed along with conversion. For better visualization, a Graphical User Interface is made to display the waveform graph, spectrogram graph and speech emotion prediction graph of a given speech audio. Meanwhile, the effect of voice activity detection algorithm is analyzed in this study. The timestamps for emotion annotation can be obtained by the result of voice activity detection.

IntroductoryThe evolution of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has resulted in the emergence of many new variant lineages that have exacerbated the COVID-19 pandemic. Some of those variants were designated as variants of concern/interest (VOC/VOI) by national or international authorities based on many factors including their potential impact on vaccines. To ascertain and rank the risk of VOCs and VOIs, we analyzed their ability to escape from vaccine-induced antibodies. The variants showed differential reductions in neutralization and replication titers by post-vaccination sera. Although the Omicron variant showed the most escape from neutralization, sera collected after a third dose of vaccine (booster sera) retained moderate neutralizing activity against that variant. Therefore, vaccination remains the most effective strategy to combat the COVID-19 pandemic.

Background Acute respiratory distress syndrome (ARDS) is a common cause of death in ICU patients and its underlying mechanism remains unclear, which leads to its high mortality rate. This study aimed to identify candidate genes potentially implicating in the pathogenesis of ARDS and provide novel therapeutic targets. Methods Using bioinformatics tools, we searched for differentially expressed genes (DEGs) in an ARDS microarray dataset downloaded from the Gene Expression Omnibus (GEO) database. Afterwards, functional enrichment analysis of GO, KEGG, GSEA and WGCNA were carried out to investigate the potential involvement of these DEGs. Moreover, the Protein–protein interaction (PPI) network was constructed and molecular complexes and hub genes were identified, followed by prognosis analysis of the hub genes. Further, we performed qRT-PCR, Western Blot and flow cytometry analysis to detect candidate genes of CCR2 and FPR3 in macrophage model of LPS-induced ARDS and primary alveolar macrophages(AMs). Macrophage chemotaxis was evaluated using Transwell assay. Results DEGs mainly involved in myeloid leukocyte activation, cell chemotaxis, adenylate cyclase-modulating G protein-coupled receptor signaling pathway and cytokine-cytokine receptor interaction. Basing on the constructed PPI network, we identified five molecular complexes and 10 hub genes potentially participating in the pathogenesis of ARDS. It was observed that candidate genes of CCR2 and FPR3 were significantly over-expressed in primary alveolar macrophages from ARDS patients and macrophgae model of LPS-induced ARDS. Moreover, in vitro transwell assay demonstrated that CCR2 and FPR3 down-regulation, respectively, inhibited LPS-triggered macrophage chemotaxis toward CCL2. Finally, a positive correlation between FPR3 and CCR2 expression was confirmed using pearson correlation analysis and Western Blot assay. Conclusions Our study identified CCR2 and FPR3 as the candidate genes which can promote macrophage chemotaxis through a possible interaction between FPR3 and CCL2/CCR2 axis and provided novel insights into ARDS pathogenesis.

The results of this study showed that the single-nucleotide polymorphism (SNP) sites of the PRL and PRLR genes have a certain association with the milk production performance, body size and cashmere performance of Liaoning cashmere goats (LCGs). Through our designed experiment, the potential SNPs of LCG were detected by sequence alignment, and two SNPs were found on two genes. The CC genotype of the PRL gene is the dominant genotype among the three genotypes. The GG genotype of the PRLR gene is the dominant genotype among the two genotypes. At the same time, the two genotypes also have good performance in cashmere production and body size. Through the screening of haplotype combination, the milk fat rate > 7.6 %, the milk protein rate > 5.6 %, the milk somatic cell number < 1500 × 103 mL−1, the cashmere fineness < 15.75 µm, the chest girth > 105 cm, the chest depth > 33 cm, and the waist height > 67.5 cm are considered as screening indexes for comprehensive production performance of Liaoning cashmere goats. It is concluded that the GCGC type is the dominant haplotype combination. According to our research data, we found that the biological indicators of Liaoning cashmere goat milk are higher than the national standards, so we think it is very significant to study the milk production performance of our experiment. Further research can be done on goat milk production and body conformation traits around PRL gene and PRLR gene.

Influenza A virus is a negative-sense RNA virus with a segmented genome consisting of eight RNA segments. Avian influenza A virus (AIV) primarily infects avian hosts and sporadically infects mammals, which can lead to adaptation to new species. Next-generation sequencing (NGS) of emerging AIV genomes extracted from respiratory samples collected on sequential days from animal models and clinical patients enables analysis of the emergence of evolutionary variants within the virus population over time. However, obtaining codon complete AIV genome at a sufficient coverage depth for nucleotide variant calling remains a challenge, especially from post-inoculation respiratory samples collected at late time points that have low viral titers. In this study, nasal wash samples from ferrets inoculated with different subtypes of AIV were collected on various days post-inoculation. Each nasal wash sample was aliquoted and extracted using five commercially available nucleic acid extraction methods. Extracted influenza virus RNA was amplified and NGS conducted using Illumina Mi-Seq. For each nasal wash sample, completeness of AIV genome segments and coverage depth were compared among five extraction methods. Nucleic acids extracted by MagNA pure compact RNA isolation consistently yielded codon complete sequences for all eight genome segments at the required coverage depth at each time point sampled. The study revealed that DNase treatment was critical to the amplification of influenza genome segments and the downstream success of codon complete NGS from nasal wash samples. The findings from this study can be applied to improve NGS of influenza and other RNA viruses that infect the respiratory tract and are collected from respiratory samples.

Recombinant SHFV infectious cDNA clones expressing a foreign gene from an additional sg mRNA were constructed. Two 3′ genomic region sites, between ORF4′ and ORF2b and between ORF4 and ORF5, were utilized for insertion of the myxoma M013 gene with a C-terminal V5 tag followed by one of the three inserted transcription regulatory sequences (TRS), TRS2′, TRS4′ or TRS7. M013 insertion at the ORF4′/ORF2b site but not the ORF4/ORF5 site generated progeny virus but only the recombinant virus with an inserted TRS2′ retained the entire M013 gene through passage four. Insertion of an auto-fluorescent protein gene, iLOV, with an inserted TRS2′ at the ORF4′/ORF2b site, generated viable progeny virus. iLOV expression was maintained through passage eight. Although regulation of SHFV subgenomic RNA synthesis is complex, the ORF4′/ORF2b site, which is located between the two sets of minor structural proteins, is able to tolerate foreign gene insertion.

Under the background of the continuous expansion of the scale and capacity of power grid and the sharp increase of information measurement points, the amount of data that needs to be analyzed and processed is growing geometrically. It is urgent to carry out the research of massive information intelligent collection and efficient computing processing scheme and storage scheme. In this paper, combined with the actual demand of monitoring data access and lossless compression method in engineering practice, we build Hadoop cloud cluster platform and hive distributed data warehouse, take analog measurement data set as cluster compression object, realize the effective compression and storage of measurement information, which has important theoretical significance and practical value for solving power grid large data processing.