Deep Learning for Real-Time Crime Forecasting and Its Ternarization

Real-time crime forecasting is important. However, accurate prediction of when and where the next crime will happen is difficult. No known physical model provides a reasonable approximation to such a complex system. Historical crime data are sparse in both space and time and the signal of interests is weak. In this work, the authors first present a proper representation of crime data. The authors then adapt the spatial temporal residual network on the well represented data to predict the distribution of crime in Los Angeles at the scale of hours in neighborhood-sized parcels. These experiments as well as comparisons with several existing approaches to prediction demonstrate the superiority of the proposed model in terms of accuracy. Finally, the authors present a ternarization technique to address the resource consumption issue for its deployment in real world. This work is an extension of our short conference proceeding paper [Wang, B., Zhang, D., Zhang, D. H., et al., Deep learning for real time Crime forecasting, 2017, arXiv: 1707.03340].     

Global regularity for vortex patches

We present a proof of Chemin's [4] result which states that the boundary of a vortex patch remains smooth for all time if it is initially smooth.Partially supported by a National Science Foundation Postdoctoral FellowshipPartially supported by the National Science Foundation     

Noncovalent cell surface engineering: incorporation of bioactive synthetic glycopolymers into cellular membranes

The controlled addition of structurally defined components to live cell membranes can facilitate the molecular level analysis of cell surface phenomena. Here we demonstrate that cell surfaces can be engineered to display synthetic bioactive polymers at defined densities by exogenous membrane insertion. The polymers were designed to mimic native cell-surface mucin glycoproteins, which are defined by their dense glycosylation patterns and rod-like structures. End-functionalization with a hydrophobic anchor permitted incorporation into the membranes of live cultured cells. We probed the dynamic behavior of cell-bound glycopolymers bearing various hydrophobic anchors and glycan structures using fluorescence correlation spectroscopy (FCS). Their diffusion properties mirrored those of many natural membrane-associated biomolecules. Furthermore, the membrane-bound glycopolymers were internalized into early endosomes similarly to endogenous membrane components and were capable of specific interactions with protein receptors. This system provides a platform to study cell-surface phenomena with a degree of chemical control that cannot be achieved using conventional biological tools.     

Synthetic analogues of glycosylphosphatidylinositol-anchored proteins and their behavior in supported lipid bilayers

Positioned at the C-terminus of many eukaryotic proteins, the glycosylphosphatidylinositol (GPI) anchor is a posttranslational modification that anchors the modified proteins in the outer leaflet of the plasma membrane. GPI-anchored proteins play vital roles in signal transduction, the vertebrate immune response, and the pathobiology of trypanosomal parasites. While many GPI-anchored proteins have been characterized, the biological functions of the GPI anchor have yet to be elucidated at a molecular level. We synthesized a series of GPI-protein analogues bearing modified anchor structures that were designed to dissect the contribution of various glycan components to the GPI-protein's membrane behavior. These anchor analogues were similar in length to native GPI anchors and included mimics of the native structure's three domains. A combination of expressed protein ligation and native chemical ligation was used to attach these analogues to the green fluorescent protein (GFP). These modified GFPs were incorporated in supported lipid bilayers, and their mobilities were analyzed using fluorescence correlation spectroscopy. The data from these experiments suggest that the GPI anchor is more than a simple membrane-anchoring device; it also may prevent transient interactions between the attached protein and the underlying lipid bilayer, thereby permitting rapid diffusion in the bilayer. The ability to generate chemically defined analogues of GPI-anchored proteins is an important step toward elucidating the molecular functions of this interesting post-translational modification.     

Synthesis and Microcontact Printing of Dual End‐Functionalized Mucin‐like Glycopolymers for Microarray Applications

Click to view : Glycopolymers can be used to display glycans on microarrays in native‐like architectures. The structurally uniform alkyne‐terminated mucin mimetic glycopolymers (see picture; TR=fluorophore) were printed on azide‐functionalized chips by microcontact printing in the presence of a copper catalyst. The surface‐bound glycopolymers bind lectins in a ligand‐specific manner.

     

Metabolic Labeling of Sialic Acids in Living Animals with Alkynyl Sugars

Sialome sweet sialome : As sialic acids are involved in many host–pathogen recognition events and are markers of embryonic and malignant tissues, there is great interest in methods for the enrichment and identification of sialylated glycoproteins from complex tissues. Now N‐(4‐pentynoyl)mannosamine can be used to metabolically label sialylated glycoproteins in living animals, enabling future identification of new biomarkers.

     

Fusion around the barrier for 7Li+12C

Fusion cross-sections for the ⁷Li + ¹²C reaction have been measured at energies above the Coulomb barrier by the direct detection of evaporation residues. The heavy evaporation residues with energies below 3 MeV could not be separated out from the α-particles in the spectrum and hence their contribution was estimated using statistical model calculations. The present work indicates that suppression of fusion cross-sections due to the breakup of ⁷Li may not be significant for ⁷Li + ¹²C reaction at energies around the barrier.