A neural network algorithm for the multiple traveling salesmen problem

We developed an efficient neural network algorithm for solving the Multiple Traveling Salesmen Problem (MTSP). A new transformation of the N-city M-salesmen MTSP to the standard Traveling Salesmen Problem (TSP) is introduced. The transformed problem is represented by an expanded version of Hopfield-Tank's neuromorphic city-position map with (N + M-1)-cities and a single fictitious salesmen. The dynamic model associated with the problem is based on the Basic Differential Multiplier Method (BDMM) [26] which evaluates Lagrange multipliers simultaneously with the problem's state variables. The algorithm was successfully tested on many problems with up to 30 cities and five salesmen. In all test cases, the algorithm always converged to valid solutions. The great advantage of this kind of algorithm is that it can provide solutions to complex decision making problems directly by solving a system of ordinary differential equations. No learning steps, logical if statements or adjusting of parameters are required during the computation. The algorithm can therefore be implemented in hardware to solve complex constraint satisfaction problems such as the MTSP at the speed of analog silicon VLSI devices or possibly future optical neural computers.     

On auxospore formation inCaloneis and the nature ofAmphiraphia (Bacillariophyta)

Amphiraphia Chen & Zhu, together with theAmphiraphiaceae andAmphiraphidales, should be abandoned, sinceAmphiraphia cells are the heterovalvar initial cells ofCaloneis. WhenCaloneis reproduces sexually two cells pair and become surrounded by a two-layered capsule of mucilage. Each cell produces two gametes and these become rearranged within the gametangial frustule before plasmogamy. The gametes are amoeboid and fusion is isogamous. Following plasmogamy the zygote contracts, becomes ellipsoidal, and lays down a primary perizonial band. This is a complete, wide hoop, while subsequent perizonial bands are narrow and open.     

The reassociation of factor Va from its isolated subunits

Factor Va is an essential cofactor for the activation of prothrombin catalyzed by factor Xa. The cofactor is a heterodimer composed of a light chain and a heavy chain that are associated noncovalently in the presence of divalent metal ions. The kinetics of the formation of factor Va from the isolated and separated subunits was examined by the time-dependent regain in cofactor activity using direct assays of prothrombin activation catalyzed by prothrombinase. The rate of reassociation at saturating concentrations of calcium ions was slow with a strong temperature dependence. The product of the association reaction was indistinguishable from native factor Va on the basis of activity. The second order rate constant for the process at 37 degrees C in the presence of 2 mM CaCl2 was 1.58 X 10(5) M-1.min-1. Manganese ion increased the rate of regain of activity without influencing the extent of the reaction. The previous identification of a single reactive sulfhydryl in each subunit of factor Va permitted the modification of the separated subunits with sulfhydryl-directed fluorophores. Subunit reassociation was directly measured by fluorescence energy transfer using light chain modified with 6-acryloyl-2-dimethylaminonaphthalene (fluorescence donor) and heavy chain modified with fluorescein 5-maleimide (fluorescence acceptor). Fluorescence measurements indicate that the heavy and light chains associate tightly (Kd = 5.9 x 10(-9) M) and reversibly with a stoichiometry of 1:1. The dissociation of the subunits from the cofactor is first order with a rate constant of 1.03 X 10(-3) min-1. These interpretations were confirmed by physical measurements of subunit reassociation by sedimentation velocity studies.     

Subunit structure of the galactose and N-acetyl-D-galactosamine-inhibitable adherence lectin of Entamoeba histolytica

The galactose and N-acetyl-D-galactosamine-inhibitable adherence lectin of Entamoeba histolytica is a cell surface protein which mediates parasite adherence to human colonic mucus, colonic epithelial cells, and other target cells. The amebic lectin was purified in 100-micrograms quantities from detergent-solubilized trophozoites by monoclonal antibody affinity chromatography. The adherence lectin was purified 500-fold as judged by radioimmunoassay. The nonreduced lectin had a molecular mass of 260 kDa on sodium dodecyl sulfate-polyacrylamide gel electrophoresis and an isoelectric point of pH 6.2. The amebic lectin reduced with beta-mercaptoethanol consisted of 170- and 35-kDa subunits. Both subunits could be labeled on the cell surface with 125I, and both were metabolically labeled with [3H]glucosamine. The amino termini of the subunits had unique amino acid sequences, and polyclonal antisera to the heavy subunit did not cross-react with the light subunit. The yield of phenylthiohydantoin derivatives from the second and third positions in the sequence of the heavy and light subunits gave a molar ratio of one 170- to one 35-kDa subunit. Antibodies directed to the heavy subunit inhibited amebic adherence to Chinese hamster ovary cells by 100%, suggesting that the heavy subunit is predominantly responsible for mediating amebic adherence.     

The activation of bovine protein C by factor Xa

A complex composed of factor Xa and phospholipid vesicles assembled in the presence of calcium ions catalyzes a discrete cleavage of the heavy chain of bovine protein C that is indistinguishable from that produced by thrombin as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. This cleavage generates an active site capable of hydrolyzing small substrates and inactivating factor Va function in the prothrombinase complex. Activation of protein C by factor Xa requires both calcium ions and phospholipid vesicles and proceeds at a rate an order of magnitude greater than that observed for alpha-thrombin in solution. gamma-Carboxyglutamic acid-domainless protein C is not activated by factor Xa, consistent with the requirement for phospholipid and distinguishing this reaction from protein C activation by thrombin. Thrombomodulin serves as a cofactor for the factor Xa-catalyzed reaction, forming a 1:1 complex with factor Xa (apparent Kd = 5.7 X 10(-10) M) and stimulating the saturated rate of protein C activation by factor Xa (kcat = 149 min-1) to levels comparable with the thrombin-thrombomodulin complex. Protein C activation by factor Xa is not inhibited by the specific thrombin inhibitor dansyl-N-(3-ethyl-1,5-pentanediyl)amide but is inhibited by antithrombin III, tripeptide-chloromethyl ketones, and the monoclonal antibody alpha-BFX-2b that is highly specific for factor Xa. These data indicate that thrombomodulin is promiscuous in its role as a cofactor and suggest the existence of an alternative pathway for protein C activation in vivo.     

Optical verification of a technique for in situ ultrasonic measurement of articular cartilage thickness

Several investigators have used pulse-echo ultrasonics to measure the thickness of articular cartilage in situ. The underlying assumption in all measurements was that the second reflection used in thickness calculations was from the calcified-cartilage/cartilage boundary (tidemark). To investigate this assumption, the thickness of 24 cartilage plugs excised from a human femoral head was measured both ultrasonically and optically. Measurements established that the second reflection was from the tidemark and validated the ultrasonic technique as a method of mapping the thickness distribution of articular cartilage in synovial joints in situ.     

Regulation of collagenase and collagenase mRNA production in early- and late-passage human diploid fibroblasts

The levels of collagenase and collagenase mRNA produced by early-passage (less than 40% of lifespan completed) and late-passage (greater than 80% of lifespan completed) cultures of human fibroblasts were analyzed. The constitutive levels of collagenase and collagenase mRNA produced by the late-passage cultures were 10-30 x greater than the levels observed in similarly treated early-passage cultures. Immunofluorescence analysis established that the percentage of collagenase-positive cells was also greater (77% vs. 4%) in the late-passage cultures. To determine whether the difference in collagenase production resulted from cell-derived regulatory factors, collagenase production was examined in cultures plated onto substrates coated with fibroblast extracellular matrix (ECM). Collagenase and collagenase mRNA production was enhanced in both types of cultures, although amounts produced by ECM-induced early-passage cultures was significantly less than that produced by similarly treated late-passage cultures. Collagen-coated substrates also induced collagenase synthesis.     

Insights into the complex association of bovine factor Va with acidic-lipid-containing synthetic membranes.

The mechanism of binding of blood coagulation cofactor factor Va to acidic-lipid-containing membranes has been addressed. Binding isotherms were generated at room temperature using the change in fluorescence anisotropy of pyrene-labeled bovine factor Va to detect binding to sonicated membrane vesicles containing either bovine brain phosphatidylserine (PS) or 1,2-dioleoyl-3-sn-phosphatidylglycerol (DOPG) in combination with 1-palmitoyl-2-oleoyl-3-sn-phosphatidylcholine (POPC). The composition of the membranes was varied from 0 to 40 mol% for PS/POPC and from 0 to 65 mol % for DOPG/POPC membranes. Fitting the data to a classical Langmuir adsorption model yielded estimates of the dissociation constant (Kd) and the stoichiometry of binding. The values of Kd defined in this way displayed a maximum at low acidic lipid content but were nearly constant at intermediate to high fractions of acidic lipid. Fitting the binding isotherms to a two-process binding model (nonspecific adsorption in addition to binding of acidic lipids to sites on the protein) suggested a significant acidic-lipid-independent binding affinity in addition to occupancy of three protein sites that bind PS in preference to DOPG. Both analyses indicated that interaction of factor Va with an acidic-lipid-containing membrane is much more complex than those of factor Xa or prothrombin. Furthermore, a change in the conformation of bound pyrene-labeled factor Va with surface concentration of acidic lipid was implied by variation of both the saturating fluorescence anisotropy and the binding parameters with the acidic lipid content of the membrane. Finally, the results cannot support the contention that binding occurs through nonspecific adsorption to a patch or domain of acidic lipids in the membrane. Factor Va is suggested to associate with membranes by a complex process that includes both acidic-lipid-specific and acidic-lipid-independent sites and a protein structure change induced by occupancy of acidic-lipid-specific sites on the factor Va molecule.