Search

Precision positioning is a very important task for automatic assembly and inspection in the manufacturing process. The conventional image processing for image alignment has relied on template matching, which is computationally intensive for objects in arbitrary locations and orientations. In this article, we propose the deep neural network regressors for fast and accurate image alignment. They are especially applied to the positioning of printed circuit boards (PCBs). The simple multilayer perceptron (MLP), the convolutional neural network (CNN), and the CNN models incorporated with support vector regression (SVR) are proposed and evaluated for the PCB positioning task. The proposed deep neural networks require only one single reference sample with a manually marked template window. All training images and the ground-truth geometric parameters are automatically generated for the model training. The effect of illumination changes and the strategies to cope with lighting variations are analyzed and proposed for robust positioning. Experimental results indicate that the proposed regressors can achieve a subpixel accuracy in translation and yield a rotation error less than 1° with 1-ms evaluation time for PCB positioning.

In dicotyledonous plants, xyloglucan (XyG) is the most abundant hemicellulose of the primary cell wall. The enzymes involved in XyG biosynthesis have been identified through reverse-genetics and activity was characterized by heterologous expression. Currently, there is no information on the atomic structures or amino acids involved in activity or substrate binding of any of the Golgi-localized XyG biosynthetic enzymes. A homology model of the xyloglucan xylosyltransferase 2 (XXT2) catalytic domain was built on the basis of the crystal structure of A64Rp. Molecular dynamics simulations revealed that the homology model retains the glycosyltransferase (GT)-A fold of the template structure used to build the homology model indicating that XXT2 likely has a GT-A fold. According to the XXT2 homology model, six amino acids (Phe204, Lys207, Asp228, Ser229, Asp230, His378) were selected and their contribution in catalytic activity was investigated. Site-directed mutagenesis studies show that Asp228, Asp230 and His378 are critical for XXT2 activity and are predicted to be involved in coordination of manganese ion. Lys207 was also found to be critical for protein activity and the homology model indicates a critical role in substrate binding. Additionally, Phe204 mutants have less of an impact on XXT2 activity with the largest effect when replaced with a polar residue. This is the first study that investigates the amino acids involved in substrate binding of the XyG-synthesizing xylosyltransferases and contributes to the understanding of the mechanisms of polysaccharide-synthesizing GTs and XyG biosynthesis.

Xyloglucan is an important hemicellulosic polysaccharide in dicot primary cell walls. Most of the enzymes involved in xyloglucan synthesis have been identified. However, many important details of its synthesis in vivo remain unknown. The roles of three genes encoding xylosyltransferases participating in xyloglucan biosynthesis in Arabidopsis (Arabidopsis thaliana) were further investigated using reverse genetic, biochemical, and immunological approaches. New double mutants (xxt1 xxt5 and xxt2 xxt5) and a triple mutant (xxt1 xxt2 xxt5) were generated, characterized, and compared with three single mutants and the xxt1 xxt2 double mutant that had been isolated previously. Antibody-based glycome profiling was applied in combination with chemical and immunohistochemical analyses for these characterizations. From the combined data, we conclude that XXT1 and XXT2 are responsible for the bulk of the xylosylation of the glucan backbone, and at least one of these proteins must be present and active for xyloglucan to be made. XXT5 plays a significant but as yet uncharacterized role in this process. The glycome profiling data demonstrate that the lack of detectable xyloglucan does not cause significant compensatory changes in other polysaccharides, although changes in nonxyloglucan polysaccharide amounts cannot be ruled out. Structural rearrangements of the polysaccharide network appear responsible for maintaining wall integrity in the absence of xyloglucan, thereby allowing nearly normal plant growth in plants lacking xyloglucan. Finally, results from immunohistochemical studies, combined with known information about expression patterns of the three genes, suggest that different combinations of xylosyltransferases contribute differently to xyloglucan biosynthesis in the various cell types found in stems, roots, and hypocotyls.

Arabidopsis thaliana xyloglucan has an XXXG structure, with branches of xylosyl residues, β-D-galacosyl-(1,2)-α-d-xylosyl motifs and fucosylated β-D-galactosyl-(1,2)-α-D-xylosyl motifs. Most of the enzymes involved in xyloglucan biosynthesis in Arabidopsis have been identified, including the glucan synthase CSLC4 (cellulose synthase-like C4), three xylosyltransferases (XXT1, XXT2 and XXT5), two galactosyltransferases (MUR3 and XLT2) and the fucosyltransferase FUT1. The XXTs and CSLC4 form homo- and heterocomplexes and were proposed to co-localize in the same complex, but the organization of the other xyloglucan-synthesizing enzymes remains unclear. Here we investigate whether the glycosyltransferases MUR3, XLT2 and FUT1 interact with the XXT-CSLC4 complexes in the Arabidopsis Golgi. We used co-immunoprecipitation and bimolecular fluorescence complementation, with signal quantification by flow cytometry, to demonstrate that CSLC4 interacts with MUR3, XLT2 and FUT1. FUT1 forms homocomplexes and interacts with MUR3, XLT2, XXT2 and XXT5. XLT2 interacts with XXT2 and XXT5, but MUR3 does not. Co-immunoprecipitation assays showed that FUT1 forms a homocomplex through disulfide bonds, and formation of the heterocomplexes does not involve covalent interactions. In vitro pull-down assays indicated that interactions in the FUT1-MUR3 and FUT1-XXT2 complexes occur through the protein catalytic domains. We propose that enzymes involved in xyloglucan biosynthesis are functionally organized in multiprotein complexes localized in the Golgi.

Xyloglucan is the major hemicellulosic polysaccharide in the primary cell walls of most vascular dicotyledonous plants and has important structural and physiological functions in plant growth and development. In Arabidopsis (Arabidopsis thaliana), the 1,4-β-glucan synthase, Cellulose Synthase-Like C4 (CSLC4), and three xylosyltransferases, XXT1, XXT2, and XXT5, act in the Golgi to form the xylosylated glucan backbone during xyloglucan biosynthesis. However, the functional organization of these enzymes in the Golgi membrane is currently unknown. In this study, we used bimolecular fluorescence complementation and in vitro pull-down assays to investigate the supramolecular organization of the CSLC4, XXT1, XXT2, and XXT5 proteins in Arabidopsis protoplasts. Quantification of bimolecular fluorescence complementation fluorescence by flow cytometry allowed us to perform competition assays that demonstrated the high probability of protein-protein complex formation in vivo and revealed differences in the abilities of these proteins to form multiprotein complexes. Results of in vitro pull-down assays using recombinant proteins confirmed that the physical interactions among XXTs occur through their catalytic domains. Additionally, coimmunoprecipitation of XXT2YFP and XXT5HA proteins from Arabidopsis protoplasts indicated that while the formation of the XXT2-XXT2 homocomplex involves disulfide bonds, the formation of the XXT2-XXT5 heterocomplex does not involve covalent interactions. The combined data allow us to propose that the proteins involved in xyloglucan biosynthesis function in a multiprotein complex composed of at least two homocomplexes, CSLC4-CSLC4 and XXT2-XXT2, and three heterocomplexes, XXT2-XXT5, XXT1-XXT2, and XXT5-CSLC4.

Given an LB-compact floorplan, a 3D block-level thermal model is firstly proposed to calculate the temperature of each circuit block in reasonable time. Furthermore, based on the temperature calculation in the proposed 3D block-level thermal model and the final floorplan region, an iterative approach is proposed to reduce the final floorplan temperature by inserting or redistribution the feasible white space. The experimental results show that our proposed iterative approach obtains very promising temperature reduction in reasonable CPU time for MCNC benchmarks.

Given a 3D non-slicing floorplan, based on the property of packing-driven ordering in the floorplan, all the blocks can be ordered as a block list. Furthermore, according to the maintenance of a dynamic sliceable floorplan, any 3D non-slicing floorplan can be transformed into a 3D slicing floorplan by shifting the related blocks to minimize the final floorplan volume. The experimental results show the proposed approach is more effective than the modified FTP-1 algorithm. Clearly, the proposed approach can serve as a post-processing phase for other non-slicing floorplan algorithms.