Accuracy of suboptimal solutions to kernel principal component analysis

For Principal Component Analysis in Reproducing Kernel Hilbert Spaces (KPCA), optimization over sets containing only linear combinations of all n-tuples of kernel functions is investigated, where n is a positive integer smaller than the number of data. Upper bounds on the accuracy in approximating the optimal solution, achievable without restrictions on the number of kernel functions, are derived. The rates of decrease of the upper bounds for increasing number n of kernel functions are given by the summation of two terms, one proportional to n ^−1/2 and the other to n ⁻¹, and depend on the maximum eigenvalue of the Gram matrix of the kernel with respect to the data. Primal and dual formulations of KPCA are considered. The estimates provide insights into the effectiveness of sparse KPCA techniques, aimed at reducing the computational costs of expansions in terms of kernel units.     

A complete treatment of a steady-state four-level model for the interpretation of OH laser-induced fluorescence measurements in atmospheric pressure flames

In order to obtain OH concentration measurements in atmospheric pressure flames from saturated fluorescence signal, we have developed a stationary four-level model. A general relationship between the measured upper state population and the total number density has been derived. The effects of the imbalance between the rotational transfer rates in the upper and ground states is analyzed. We show that in the case of comparable rotational rates, the results of our treatment are similar to the ones obtained with the balanced cross rate model. A set of complementary data which should facilitate absolute OH concentration measurements is also pointed out.     

Use of 2,2-dimethoxypropane for the direct gas chromatographic-mass spectrometric determination of some organic compounds in water

The well known reaction between 2,2-dimethoxypropane and water allows for the conversion of an aqueous into an organic solution ready to be injected directly into a gas chromatographic-mass spectrometric (GC-MS) system. Only time, temperature and water/ketal ratio were the parameters studied among those which influence the equilibrium reaction. The reaction environment is not suitable for all compounds: acids, esters, alcohols, amines, ketones and phenols may react with methanol, acetone and ketal owing to Amberlyst being present as catalyst. This method is proposed for the GC-MS analysis of aqueous solutions containing hydrocarbons, halogenated hydrocarbons and ethers. The determination of some pesticides in water is reported.     

A boundary-integral equation method for free surface viscous flows

Summary A boundary integral equation method is proposed for the solution of viscous recirculating flows with free surfaces. In particular the method is applied to thermocapillary convection and to drop formation, both in micro-gravity conditions, the latter to test its capability to handle real unsteady problems.The presence of non linear terms in Navier-Stokes equations leads to a volume integral, which has to be approximated by a linearization procedure.Several numerical results for thermocapillary flows, both with fixed and moving free surface, are discussed in comparison with previously obtained finite difference solutions. Some preliminary results, and in particular the time evolution of the free surface shape, are also presented for the drop formation problem. Only plane two dimensional fields are considered for both problems.

---

Sommario Si propone un metodo basato sulla soluzione di equazioni integrali di contorno per flussi viscosi con superficie libera. Tale metodo è applicato allo studio della convezione termocapillare ed al processo di formazione di una goccia, entrambi in condizioni di microgravità. La presenza dei termini non lineari nell'equazione di Navier-Stokes comporta un integrale di volume che viene approssimato mediante un processo di linearizzazione.Risultati numerici per flussi termocapillari con superficie libera sia fissa che mobile sono confrontati con altri ottenuti in precedenza con un metodo alle differenze finite. Si presentano inoltre alcuni risultati preliminari sul problema della formazione della goccia ed in particolare l'evoluzione nel tempo della configurazione geometrica della superficie libera. Nei due casi si analizzano solo campi bidimensionali.

---

---

Presented at the VII National Conference AIDAA, Naples, September 1983.

---

In leave of absence from Tianjin University, China.     

SRK equation of state: Predicting binary interaction parameters of hydrocarbons and related compounds

The parameter mixing rules of the Soave–Redlich–Kwong (SRK) equation of state are rewritten as Huron–Vidal mixing rules, where infinite-pressure activity coefficients are predicted by group contributions. Alkanes are treated as composed by one group type and aromatics by two types, aliphatic and aromatic. Hydrocarbon mixtures can be treated using one universal interaction parameter. Light compounds like methane, N₂, CO₂, H₂S, etc. are treated as separate groups; each one requires a pair of parameters for its interactions with aliphatic and aromatic groups. Group interaction parameters were determined from experimental VLE data. From them, binary interaction constants of the classical quadratic mixing rules can directly be derived.     

FTIR microspectroscopy of melanocytic skin lesions: a preliminary study

Fourier transform infrared (FTIR) microspectroscopy has been employed to investigate benign (ordinary dermal and Reed nevi), dysplastic and malignant (invasive melanoma) skin lesions through the analysis of spectral changes of melanocytes as well as in the evaluation of the presence of melanin. Hierarchical cluster analysis and principal component analysis led to a satisfactory separation of malignant from dysplastic and normal melanocytes. Also, on enlarging the clustering with spectra from Reed and dermal nevi, the multivariate analysis segregated well the spectral data into discrete clusters, allowing the obtaining of reliable average spectra for analysis at the molecular level of the main groups or components responsible for the biological and biochemical changes. The most significant spectral characteristics appear to be related to differences in secondary protein structures, in nucleic acid conformation, in intra- and intermolecular bonding. In all cases, supervised and unsupervised spectral analyses resulted in satisfactory agreement with histopathological findings.     

Effects of Lactobacillus rhamnosus on zebrafish oocyte maturation: an FTIR imaging and biochemical analysis

The aim of this study was to verify the effects of probiotic Lactobacillus rhamnosus on zebrafish oocyte maturation using FPA (focal plane array) FTIR imaging together with specific biochemical assays (SDS-PAGE, real-time PCR and enzymatic assay). Oocyte growth is prevalently due to a vitellogenic process which consists of the hepatic synthesis of vitellogenin and its selective uptake during maturation. The administration of L. rhamnosus IMC 501 for 10 days induced chemical changes to oocyte composition, promoting the maturation process. Some interesting biochemical features, linked to protein secondary structure (amide I band) and to phospholipidic and glucidic patterns, were detailed by vibrational analysis. The spectroscopic results were supported by the early increase of the lysosomal enzyme involved in the final oocyte maturation, the cathepsin L. This enzyme increases during follicle maturation, with the highest levels in class IV oocytes. In treated females, class III oocytes showed higher cathepsin L gene expression and enzymatic activity, with levels comparable to class IV oocytes isolated from controls; this can be related to the proteolytic cleavage of the higher molecular mass yolk protein components, as evidenced by SDS-PAGE.     

Analysis of trace degradation products (decarboxylated diastereoisomers) of S-adenosylmethionine by electrophoresis in capillaries with cationic coatings (N-methylpolyvinylpyridinium or divalent barium)

Commercial preparations of S‐adenosylmethionine (SAM) when analyzed in uncoated capillaries show a minute impurity believed to be decarboxylated (dc) SAM. By using two types of cationic coatings, thus reducing the electro‐endo‐osmotic flow (EOF), it was possible to separate this impurity into two diastereoisomers of dcSAM. The coatings evaluated for this purpose were: (i) N‐methylpolyvinylpyridinium, used under reversed EOF at acidic conditions (pH 4.0) and (ii) deposition of divalent barium at alkaline pH values (pH 9.4), providing reduced EOF. Under these conditions, it was possible to separate this impurity into two diastereoisomers, which by chemical synthesis were indeed proven to be dcSAM. It was further demonstrated that, in the alkylation of 5′‐methylthioadenosine by 3‐bromopropylamine in bromidric acid to dcSAM, another minute impurity was present, proven, via mass spectrometry, to consist of S‐(5′‐adenosyl)‐3‐thiopropylamine (decarboxylated and demethylated (dc‐SAH)). The LOD for the two dcSAM diastereoisomers was assessed as 17.5 μg/mL and their LOQ as 25.5 μg/mL. By the barium‐based protocol it was possible to quantify the dcSAM, present in a commercial sample of SAM, as a 0.1% impurity.     

Structural Analysis of a Helical Peptide Unfolding Pathway

The analysis of the folding mechanism in peptides adopting well‐defined secondary structure is fundamental to understand protein folding. Herein, we describe the thermal unfolding of a 15‐mer vascular endothelial growth factor mimicking α‐helical peptide (QK_L10A) through the combination of spectroscopic and computational analyses. In particular, on the basis of the temperature dependencies of QK_L10A H_α chemical shifts we show that the first phase of the thermal helix unfolding, ending at around 320 K, involves mainly the terminal regions. A second phase of the transition, ending at around 333 K, comprises the central helical region of the peptide. The determination of high‐resolution QK_L10A conformational preferences in water at 313 K allowed us to identify, at atomic resolution, one intermediate of the folding–unfolding pathway. Molecular dynamics simulations corroborate experimental observations detecting a stable central helical turn, which represents the most probable site for the helix nucleation in the folding direction. The data presented herein allows us to draw a folding–unfolding picture for the small peptide QK_L10A compatible with the nucleation–propagation model. This study, besides contributing to the basic field of peptide helix folding, is useful to gain an insight into the design of stable helical peptides, which could find applications as molecular scaffolds to target protein–protein interactions.     

Characterization of the volatile organic compounds of Italian 'Fossa' cheese by solid-phase microextraction gas chromatography/mass spectrometry

Fossa cheese is an Italian hard cheese, ripened for up to 3 months in underground pits dug into tuffaceous rock. During this period, the cheese develops a unique flavour and intense and somewhat piquant aroma. Solid-phase microextraction gas chromatography/mass spectrometry (SPME-GC/MS) was utilized to characterize the volatile organic compounds (VOCs) of Fossa cheese. A total of 75 VOCs were separated and identified; in particular, the major class of compounds found in the cheeses ripened in the pits were the esters of fatty acids. Discriminant analysis of volatile profiles allowed us to distinguish between cheeses in different stages of seasoning (60-day-old cheese and cheese ripened an additional 90 days in and out of the pits).