No abstract

With the increasing clock rate and transistor count of today 's microprocessors, power

No abstract

No abstract

SUMMARY Successful pathogen infection likely involves the suppression of general antimicrobial host defences. One Pseudomonas syringae virulence factor proposed to act in this manner is coronatine (COR), a phytotoxin believed to function as an analogue of one or more jasmonates, a family of plant growth regulators. COR biosynthetic (COR(-)) mutants of P. syringae pv. tomato strain DC3000 exhibit reduced virulence on Arabidopsis thaliana and tomato. In the present study, three genetically and biochemically defined COR(-) mutants of DC3000 were used to explore potential effects of COR and its precursors, coronafacic acid (CFA) and coronamic acid (CMA), on defence signalling pathways in A. thaliana. Inoculation with wild-type DC3000 resulted in the accumulation of several jasmonate-responsive transcripts, whereas infection with a mutant strain that accumulates CFA, which is structurally similar to methyl jasmonate (MeJA), did not. Thus, COR, but not CFA, stimulates jasmonate signalling during P. syringae infection of A. thaliana. The ability of the COR(-) mutants to grow to high levels in planta was fully restored in A. thaliana lines deficient for salicylic acid (SA) accumulation. Although the COR(-) mutants grew to high levels in SA-deficient plants, disease symptoms were reduced in these plants. Collectively, these results indicate that COR is required both for overcoming or suppressing SA-dependent defences during growth in plant tissue and for normal disease symptom development in A. thaliana.

To identify Pseudomonas syringae pv. tomato genes involved in pathogenesis, we carried out a screen for Tn5 mutants of P. syringae pv. tomato DC3000 with reduced virulence on Arabidopsis thaliana. Several mutants defining both known and novel virulence loci were identified. Six mutants contained insertions in biosynthetic genes for the phytotoxin coronatine (COR). The P. syringae pv. tomato DC3000 COR genes are chromosomally encoded and are arranged in two separate clusters, which encode enzymes responsible for the synthesis of coronafacic acid (CFA) or coronamic acid (CMA), the two defined intermediates in COR biosynthesis. High-performance liquid chromatography fractionation and exogenous feeding studies confirmed that Tn5 insertions in the cfa and cma genes disrupt CFA and CMA biosynthesis, respectively. All six COR biosynthetic mutants were significantly impaired in their ability to multiply to high levels and to elicit disease symptoms on A. thaliana plants. To assess the relative contributions of CFA, CMA, and COR in virulence, we constructed and characterized cfa6 cmaA double mutant strains. These exhibited virulence phenotypes on A. thalliana identical to those observed for the cmaA or cfa6 single mutants, suggesting that reduced virulence of these mutants on A. thaliana is caused by the absence of the intact COR toxin. This is the first study to use biochemically and genetically defined COR mutants to address the role of COR in pathogenesis.

No abstract

The widespread application of deep learning has changed the landscape of computation in the data center. In particular, personalized recommendation for content ranking is now largely accomplished leveraging deep neural networks. However, despite the importance of these models and the amount of compute cycles they consume, relatively little research attention has been devoted to systems for recommendation. To facilitate research and to advance the understanding of these workloads, this paper presents a set of real-world, productionscale DNNs for personalized recommendation coupled with relevant performance metrics for evaluation. In addition to releasing a set of open-source workloads, we conduct indepth analysis that underpins future system design and optimization for at-scale recommendation: Inference latency varies by 60% across three Intel server generations, batching and co-location of inferences can drastically improve latency-bounded throughput, and the diverse composition of recommendation models leads to different optimization strategies.Preprint. Under submission.