Design of Financial CDO Squared Transactions Using Constraint Programming

We give an approximate and often extremely fast method of building a particular kind of portfolio in finance, here called a portfolio design (PD), with applications in the credit derivatives market, for example when designing collateralised debt obligations squared (CDO²) transactions. A PD generalises a balanced incomplete block design (BIBD) and is usually harder to build. Worse, typical financial PDs are an order of magnitude larger than the largest BIBDs built so far by constraint programs, and in practice an optimisation version of the problem of building PDs has to be solved. Our method is based on embedding small designs, whose determination is itself a constraint satisfaction problem, into the original large design. Together with the detection of when a PD might be a BIBD, symmetry breaking, extended reuse of previously built PDs, and admissibility checking during search, the performance of the method becomes good enough for designing (near-)optimal CDO² transactions, with sizes common in the credit derivatives market, within minutes. For example, we optimally build a typical financial PD, which has over 10⁷⁴⁶ symmetries, in just a few minutes. The high quality of our approximate designs can be assessed by comparison with a lower bound on the optimum. Our designs sufficiently improve the currently best ones so as often to make the difference between having and not having a feasible CDO² transaction due to investor and rating-agency constraints.     

Mechanical Response of Thermoset Polymers under High Compressive Loads, 1

Summary: The present study describes the mechanical response of thermoset polymers under high compressive loads. A well‐defined free radically cured vinyl ester resin has been used and studied in six different geometries in order to determine the dependence of apparent mechanical properties on the particular size and shape of a sample. The mechanical response in compression has also been compared to the response in tensile tests. Variation of the film thickness, boundary conditions and loading conditions reveal that there is a significant effect on the mechanical performance (apparent properties) of the polymer. When the thickness‐to‐width ratio of the sample is reduced in a compression test, the friction between the sample and the compression plates proves to be of great importance. The yield stress increases dramatically when the thickness of the sample is reduced, whereas it decreases when the friction between sample and the compression plate is reduced. The creep decreases when the thickness of the material is reduced and it decreases even more due to reaction of the material surrounding the compressed part of the sample. The described test conditions and observed phenomena will be subject to simulation in Part 2 of this study.

Mechanical performance of a thermoset polymer under high compressive loads.     

Formation of aldonic acid end groups during NO2/O2 treatment and alkali cooking of wood meal

Summary The increased carbohydrate yield after NaOH cooking of wood resulting from NO₂/O₂ pretreatment is related to the oxidation of reducing sugar end groups to aldonic acid groups during both the pretreatment and the subsequent alkali cook. This leads to a suppressed end-wise degradation during the alkali cook, reflected in a lowered formation of 3-deoxyaldonic and 2-C-methylglyceric acid end groups in competing reactions.     

Radial cracks in boron fibres

The morphology of so-called radial cracks in boron fibres is studied in transverse and axial cross-sections, as well as in fracture surfaces. A formation mechanism is proposed in terms of the transverse residual stress pattern. Small pre-existing voids, (so-called proximate voids) are found to act as stress concentrators which contribute to the opening up of radial cracks in quantitative agreement with the Griffith criterion for brittle fracture.     

Copper smelting in Boliden’s Rönnskär works described

As much of the production as possible in Ronnskar works is handled during the day shift. This requires three times as large furnaces and casting machines as three shifts and involves a considerable outlay of capital, but the labor force is reduced approximately two thirds with production costs decreased to a corresponding degree.     

The use of a pharmacophore model for identification of novel ligands for the benzodiazepine binding site of the GABAA receptor

A Catalyst pharmacophore model has been developed for the benzodiazepine site within the GABA(A) receptor complex. The model is based on a pharmacophore model originally proposed by Cook and co-workers (Drug Des. Discovery 1995, 12, 193-248) and further developed by Kahnberg et al. (J. Med. Chem. 2002, 45, 4188-4201). The Catalyst pharmacophore model has been validated by using a series of flavonoids with varying affinities for the benzodiazepine receptor and has then been used as a search query in database searching with the aim of finding novel structures which have the possibility to be modified into novel lead compounds. Five of the hits from the database searching were purchased and their affinities for the benzodiazepine site of the GABA(A) receptor were determined. Two of the compounds displayed K(i) values below 10 microM. The substance showing highest potency in-vitro displayed an affinity of 121 nM making it an interesting compound for optimization. The false positive compounds (K(i) values >10 microM affinities) have been analysed in terms of conformational energy penalties and possibilities for hydrogen bond interactions. The analysis clearly demonstrates the need for post processing of Catalyst hits.     

Fully automatic macromolecular crystallography: the impact of MASSIF-1 on the optimum acquisition and quality of data

Automation is beginning to transform the way data are collected in almost all scientific disciplines. The combination of robotics and software now allows data to be collected consistently and reproducibly, eliminating human error and boredom. This approach has been applied to macromolecular crystallography at MASSIF-1, a fully automated beamline at the European Synchrotron Radiation Facility (ESRF). Considerable human effort is still dedicated to evaluating protein crystals in order to find the few crystals that diffract well or collecting hundreds of data sets to screen potential new drug candidates. The combination of ESRF-developed robotic sample handling and advanced software protocols now provides a new tool to structural biologists. Not only is the beamline used efficiently, running 24?h a day without getting tired, data collection is also performed consistently by an expert system, often better than with a human operator. In this review, we will focus on the impact this level of automation has had on the optimum acquisition of data from crystals of biological macromolecules.     

A fast and precise approach for computational saturation mutagenesis and its experimental validation by using an artificial (βα)8-barrel protein

We present a computational saturation mutagenesis protocol (CoSM) that predicts the impact on stability of all possible amino acid substitutions for a given site at an internal protein interface. CoSM is an efficient algorithm that uses a combination of rotamer libraries, side-chain flips, energy minimization, and molecular dynamics equilibration. Because CoSM considers full side-chain and backbone flexibility in the local environment of the mutated position, amino acids larger than the wild-type residue are also modeled in a proper manner. To assess the performance of CoSM, the effect of point mutations on the stability of an artificial (βα)(8)-barrel protein that has been designed from identical (βα)(4)-half barrels, was studied. In this protein, position 234(N) is a previously identified stability hot-spot that is located at the interface of the two half barrels. By using CoSM, changes in protein stability were predicted for all possible single point mutations replacing wild-type Val234(N). In parallel, the stabilities of 14 representative mutants covering all amino acid classes were experimentally determined. A linear correlation of computationally and experimentally determined energy values yielded an R(2) value of 0.90, which is statistically significant. This degree of coherence is stronger than the ones we obtained for established computational methods of mutational analysis.