PolyGLoT: A Polyhedral Loop Transformation Framework for a Graphical Dataflow Language

Polyhedral techniques for program transformation are now used in several proprietary and open source compilers. However, most of the research on polyhedral compilation has focused on imperative languages such as C, where the computation is specified in terms of statements with zero or more nested loops and other control structures around them. Graphical dataflow languages, where there is no notion of statements or a schedule specifying their relative execution order, have so far not been studied using a powerful transformation or optimization approach. The execution semantics and referential transparency of dataflow languages impose a different set of challenges. In this paper, we attempt to bridge this gap by presenting techniques that can be used to extract polyhedral representation from dataflow programs and to synthesize them from their equivalent polyhedral representation. We then describe PolyGLoT, a framework for automatic transformation of dataflow programs which we built using our techniques and other popular research tools such as Clan and Pluto. For the purpose of experimental evaluation, we used our tools to compile LabVIEW, one of the most widely used dataflow programming languages. Results show that dataflow programs transformed using our framework are able to outperform those compiled otherwise by up to a factor of seventeen, with a mean speed-up of 2.30x while running on an 8-core Intel system.

Analysis of the binding of polymyxin B to endotoxic lipid A and core glycolipid using a fluorescent displacement probe

Dansylcadaverine, a cationic fluorescent probe binds to bacterial lipopolysaccharide and lipid A, and is displaced competitively by other compounds which possess affinity toward endotoxins. The binding parameters of dansylcadaverine for lipid A were determined by Scatchard analysis to be two apparently equivalent sites with apparent dissociation constants (Kd) ranging between 16 μM to 26 μM, while that obtained for core glycolipid from Salmonella minnesota Re595 yielded a Kd of 22 μM to 28 μM with three binding sites. The Kd of polymyxin B for lipid A was computed from dansylcadaverine displacement by the method of Horovitz and Levitzki (Horovitz, A., and Levitzki, A. (1987) Proc. Natl. Acad. Sci. USA 84, 6654–6658). The applicability of this method for analyzing fluorescence data was validated by comparing the Kds of melittin for lipid A obtained by direct Scatchard analysis, and by the Horovitz-Levitzki method. The displacement of dansylcadaverine from lipid A by polymyxin B was distinctly biphasic with Kds for polymyxin B-lipid A interactions corresponding to 0.4 μM and 1.5 μM, probably resulting as a consequence of lipid A being a mixture of mono- and di-phosphoryl species. This was not observed with core glycolipid, for which the Kd for polymyxin was estimated to range from 1.1 μM to 5.8 μM. The use of dansylcadaverine as a displacement probe offers a novel and convenient method of quantitating the interactions of a wide variety of substances with lipid A.

The Escherichia coli Phosphotyrosine Proteome Relates to Core Pathways and Virulence

While phosphotyrosine modification is an established regulatory mechanism in eukaryotes, it is less well characterized in bacteria due to low prevalence. To gain insight into the extent and biological importance of tyrosine phosphorylation in Escherichia coli, we used immunoaffinity-based phosphotyrosine peptide enrichment combined with high resolution mass spectrometry analysis to comprehensively identify tyrosine phosphorylated proteins and accurately map phosphotyrosine sites. We identified a total of 512 unique phosphotyrosine sites on 342 proteins in E. coli K12 and the human pathogen enterohemorrhagic E. coli (EHEC) O157:H7, representing the largest phosphotyrosine proteome reported to date in bacteria. This large number of tyrosine phosphorylation sites allowed us to define five phosphotyrosine site motifs. Tyrosine phosphorylated proteins belong to various functional classes such as metabolism, gene expression and virulence. We demonstrate for the first time that proteins of a type III secretion system (T3SS), required for the attaching and effacing (A/E) lesion phenotype characteristic for intestinal colonization by certain EHEC strains, are tyrosine phosphorylated by bacterial kinases. Yet, A/E lesion and metabolic phenotypes were unaffected by the mutation of the two currently known tyrosine kinases, Etk and Wzc. Substantial residual tyrosine phosphorylation present in an etk wzc double mutant strongly indicated the presence of hitherto unknown tyrosine kinases in E. coli. We assess the functional importance of tyrosine phosphorylation and demonstrate that the phosphorylated tyrosine residue of the regulator SspA positively affects expression and secretion of T3SS proteins and formation of A/E lesions. Altogether, our study reveals that tyrosine phosphorylation in bacteria is more prevalent than previously recognized, and suggests the involvement of phosphotyrosine-mediated signaling in a broad range of cellular functions and virulence.

CUDA-for-clusters: a system for efficient execution of CUDA kernels on multi-core clusters

Rapid advancements in multi-core processor architectures coupled with low-cost, low-latency, high-bandwidth interconnects have made clusters of multi-core machines a common computing resource. Unfortunately, writing good parallel programs that efficiently utilize all the resources in such a cluster is still a major challenge. Various programming languages have been proposed as a solution to this problem, but are yet to be adopted widely to run performance-critical code mainly due to the relatively immature software framework and the effort involved in re-writing existing code in the new language. In this paper, we motivate and describe our initial study in exploring CUDA as a programming language for a cluster of multi-cores. We develop CUDA-For-Clusters (CFC), a framework that transparently orchestrates execution of CUDA kernels on a cluster of multi-core machines. The well-structured nature of a CUDA kernel, the growing popularity, support and stability of the CUDA software stack collectively make CUDA a good candidate to be considered as a programming language for a cluster. CFC uses a mixture of source-to-source compiler transformations, a work distribution runtime and a light-weight software distributed shared memory to manage parallel executions. Initial results on running several standard CUDA benchmark programs achieve impressive speedups of up to 7.5X on a cluster with 8 nodes, thereby opening up an interesting direction of research for further investigation.

Lipid coated mesoporous silica nanoparticles as an oral delivery system for targeting and treatment of intravacuolar Salmonella infections

Lipid coated mesoporous silica nanoparticle (L-MSN) were synthesized for oral delivery of ciprofloxacin for intracellular elimination of Salmonella pathogen. The particle size was found to be between 50-100 nm with a lipid coat of approximately 5 nm thickness. The lipid coating was achieved by sonication of liposomes with the MSN particles and evaluated by CLSMand FTIR studies. The L-MSN particles exhibited lower cytotoxicity compared to bare MSN particles. Ciprofloxacin, a fluoroquinolone antibiotic, loaded into the L-MSN particles showed enhanced antibacterial activity against free drug in in vitro assays. The lipid coat was found to aid in intravacuolar targeting of the drug cargo as observed by confocal microscopy studies. We also observed that a lower dose of antibiotic was sufficient to clear the pathogen from mice and increase their survivability using the L-MSN oral delivery system.