Genomic copies of three Xanthomonas oryzae pv. oryzae avirulence (avr) genes, avrXa7, avrXal0, and avrxa5, and four homologous genes, aB3.5, aB3.6, aB4.3, and aB4.5, were mutagenized individually or in combination to study the roles of avr genes in one component of pathogen fitness, i.e., aggressiveness or the amount of disease X. oryzae pv. oryzae causes in susceptible rice lines. These X. oryzae pv. oryzae genes are members of the highly related Xanthomonas avrBs3 gene family. Compared to the wild-type strain, X. oryzae pv. oryzae strains with mutations in avrXa7, avrxa5, and the four homologous genes caused shorter lesions on rice line IR24, which contains no resistance genes relevant to the wild-type strain. The contribution of each gene to lesion length varied, with avrXa7 contributing the most and avrXal0 showing no measurable effect on aggressiveness. The functional, plasmidborne copies of avrXa7, aB4.5, and avrxa5 restored aggressiveness only to strains with mutations in avrXa7, aB4.5, and avrxa5, respectively. Mutations in avrXa7 were not complemented by plasmids carrying any other avr gene family members. These data indicate that some, but not all, avr family members contribute to pathogen aggressiveness and that the contributions are quantitatively different. Furthermore, despite their sequence similarity, the aggressiveness functions of these gene family members are not interchangeable. The results suggest that selection and pyramiding resistance genes can be guided by the degree of fitness penalty that is empirically determined in avr gene mutations.

Question-answering (QA) on video contents is a significant challenge for achieving human-level intelligence as it involves both vision and language in real-world settings. Here we demonstrate the possibility of an AI agent performing video story QA by learning from a large amount of cartoon videos. We develop a video-story learning model, i.e. Deep Embedded Memory Networks (DEMN), to reconstruct stories from a joint scene-dialogue video stream using a latent embedding space of observed data. The video stories are stored in a long-term memory component. For a given question, an LSTM-based attention model uses the long-term memory to recall the best question-story-answer triplet by focusing on specific words containing key information. We trained the DEMN on a novel QA dataset of children's cartoon video series, Pororo. The dataset contains 16,066 scene-dialogue pairs of 20.5-hour videos, 27,328 fine-grained sentences for scene description, and 8,913 story-related QA pairs. Our experimental results show that the DEMN outperforms other QA models. This is mainly due to 1) the reconstruction of video stories in a scene-dialogue combined form that utilize the latent embedding and 2) attention. DEMN also achieved state-of-the-art results on the MovieQA benchmark.

Background We conducted a longitudinal study to evaluate changes in the clinical presentation and epidemiology of Staphylococcus aureus bacteremia (SAB) in an academic, US medical center. Methods Consecutive patients with monomicrobial SAB were enrolled from January 1995 to December 2015. Each person’s initial bloodstream S. aureus isolate was genotyped using spa typing. Clonal complexes (CCs) were assigned using Ridom StaphType software. Changes over time in both the patient and bacterial characteristics were estimated with linear regression. Associations between genotypes or clinical characteristics and complications were estimated using multivariable regression models. Results Among the 2348 eligible participants, 54.2% had an implantable, foreign body of some type. This proportion increased significantly during the 21-year study period, by 0.96% annually (P = .002), as did comorbid conditions and acquisition outside of the hospital. Rates of any metastatic complication also significantly increased, by 0.94% annually (P = .019). Among the corresponding bloodstream S. aureus isolates, spa-CC012 (multi-locus sequence type [MLST] CC30), -CC004 (MLST CC45), -CC189 (MLST CC1), and -CC084 (MLST CC15) all significantly declined during the study period, while spa-CC008 (MLST CC8) significantly increased. Patients with SAB due to spa-CC008 were significantly more likely to develop metastatic complications in general, and abscesses, septic emboli, and persistent bacteremia in particular. After adjusting for demographic, racial, and clinical variables, the USA300 variant of spa-CC008 was independently associated with metastatic complications (odds ratio 1.42; 95% confidence interval 1.02–1.99). Conclusions Systematic approaches for monitoring complications of SAB and genotyping the corresponding bloodstream isolates will help identify the emergence of hypervirulent clones and likely improve clinical management of this syndrome.

0000−0003−2426−2198] , Seong-Ho Choi 2[0000−0002−7553−6761] , Jin-Hwa Kim 3 [0000−0002−0423−0415] , and Byoung-Tak Zhang 2,4[0000−0001−9890−0389]Abstract. We propose a video story question-answering (QA) architecture, Multimodal Dual Attention Memory (MDAM). The key idea is to use a dual attention mechanism with late fusion. MDAM uses selfattention to learn the latent concepts in scene frames and captions. Given a question, MDAM uses the second attention over these latent concepts.Multimodal fusion is performed after the dual attention processes (late fusion). Using this processing pipeline, MDAM learns to infer a high-level vision-language joint representation from an abstraction of the full video content. We evaluate MDAM on PororoQA and MovieQA datasets which have large-scale QA annotations on cartoon videos and movies, respectively. For both datasets, MDAM achieves new state-of-the-art results with significant margins compared to the runner-up models. We confirm the best performance of the dual attention mechanism combined with late fusion by ablation studies. We also perform qualitative analysis by visualizing the inference mechanisms of MDAM.

Despite the identification of Acinetobacter baumannii isolates that demonstrate susceptibility to only colistin, this antimicrobial agent was not available in Korea until 2006. The present study examined the outcomes of patients with multidrug resistant (MDR) Acinetobacter species bloodstream infection and who were treated with or without colistin as part of their regimen. The colistin group was given colistin as part of therapy once colistin became available in 2006. The non-colistin group was derived from the patients who were treated with other antimicrobial regimens before 2006. Mortality within 30 days of the onset of bacteremia occurred for 11 of 31 patients in the colistin group and for 15 of 39 patients in the non-colistin group (35.5% vs 38.5%, respectively, P = 0.80). Renal dysfunction developed in 50.0% of the 20 evaluable patients in the colistin group, but in 28.6% of the 35 evaluable patients in the non-colistin group (P = 0.11). On multivariate analysis, only an Acute Physiological and Chronic Health Evaluation II score ≥ 21 was associated with mortality at 30 days. This result suggests that administering colistin, although it is the sole microbiologically appropriate agent, does not influence the 30 day mortality of patients with a MDR Acinetobacter spp. bloodstream infection.

No abstract

No abstract

The present study was conducted to screen candidate probiotic strains for anti-inflammatory activity. Initially, a nitric oxide (NO) assay was used to test selected candidate probiotic strains for anti-inflammatory activity in cultures of the murine macrophage cell line, RAW 264.7. Then, the in vitro probiotic properties of the strains, including bile tolerance, acid resistance, and growth in skim milk media, were investigated. We also performed an in vitro hydrophobicity test and an intestinal adhesion assay using Caenorhabditis elegans as a surrogate in vivo model. From our screening, we obtained 4 probiotic candidate lactic acid bacteria (LAB) strains based on their anti-inflammatory activity in lipopolysaccharide (LPS)-stimulated RAW 264.7 cell cultures and the results of the in vitro and in vivo probiotic property assessments. Molecular characterization using 16S rDNA sequencing analysis identified the 4 LAB strains as Lactobacillus plantarum. The selected L. plantarum strains (CAU1054, CAU1055, CAU1064, and CAU1106) were found to possess desirable in vitro and in vivo probiotic properties, and these strains are good candidates for further investigations in animal models and human clinical studies to elucidate the mechanisms underlying their anti-inflammatory activities.