École Nationale de la Statistique et de l'Analyse de l'Information

Abstract Data in many different fields come to practitioners through a process naturally described as functional. Although data are gathered as finite vector and may contain measurement errors, the functional form have to be taken into account. We propose a clustering procedure of such data emphasizing the functional nature of the objects. The new clustering method consists of two stages: fitting the functional data by B‐splines and partitioning the estimated model coefficients using a k ‐means algorithm. Strong consistency of the clustering method is proved and a real‐world example from food industry is given.

Despite various spectacular therapeutic properties, curcumin has low bioavailability mainly due to its poor solubility in water.

Summary Plant roots release in the rhizosphere diverse organic materials which may have different effects on soil structure. We have evaluated the effect of natural and modelled root‐released materials on soil aggregates and the biodegradation of carbon from roots in the soil. The effects of root mucilage from maize and of a modelled soluble exudate were compared with those of simple compounds (glucose, polygalacturonic acid). For all treatments, soil was amended with 2 g C kg −1 soil and incubated for 30 days at 25°C. The biodegradation of mucilage was similar to that of polygalacturonic acid, and slower than the decomposition of modelled exudates and glucose. Addition of all substrates increased the stability of aggregates, but the duration of this effect depended on the chemical nature of the material. Compared with the control, the proportion of stable aggregates after 30 days of incubation was multiplied by 3.8 for root mucilage, by 4.2 for modelled soluble exudates, by 2.5 for polygalacturonic acid and by 2.0 for glucose. The different fractions of root exudates in the rhizosphere evidently affected the aggregate stability.

Circulating concentrations of leptin in sheep correlate with body fatness and are affected by level of food intake and photoperiod. The present objective was to elucidate the short-term dynamics of leptin secretion. Frequent blood samples were taken over 48 h from 12 Soay rams after 16 weeks in short-day photoperiod (SD, 16 h darkness:8 h light) with freely available food, and then after 16 weeks in long days (16 h light:8 h darkness) with food freely available (LD) or restricted to 90% maintenance (LDR) (n=6/group). During the second 24 h of sampling, half were food deprived (n=6, SD and LD) and half had their meal times shifted (n=6, SD and LDR). A homologous RIA was developed, using antibodies raised in chicken against recombinant ovine leptin, to measure plasma concentrations. Simultaneous 24 h profiles of plasma insulin, glucose and non-esterified fatty acids (NEFA) were measured. Plasma leptin was higher in LD than SD, and in LD than LDR, associated with higher food intake, liveweight and body condition score (adiposity), but tended to be lower in LDR than SD, associated with lower food intake, liveweight and body condition score. There was no evidence for a circadian rhythm of plasma leptin, but clear evidence for post-prandial peaks of low amplitude (15-36%) 2-8 h after meals given at normal and shifted times. Complete food deprivation caused a dramatic fall in plasma leptin to basal levels within 24 h. There was a positive association of plasma leptin with plasma insulin, and negative association with NEFA, both between meals and during fasting. Thus, plasma leptin concentrations in sheep are sensitive to short-term changes in energy balance, as well as to long-term photoperiod-driven changes in food intake and adiposity.

BACKGROUND: Carbohydrate-specific IgE antibodies present on nonprimate mammalian proteins were incriminated recently in delayed meat anaphylaxis. The aim of this study was to explore whether anaphylaxis to mammalian kidney is also associated with galactose-α-1,3-galactose (αGal)-specific IgE. METHODS: Fourteen patients with anaphylaxis to pork or beef kidney underwent prick tests to meat and kidney. Some patients also underwent skin tests to Erbitux(®) (cetuximab). IgE antibodies to αGal, swine urine proteins, beef and pork meat, serum albumin proteins, cat, and rFel d 1 were measured by ImmunoCAP(®). The αGal levels were estimated in meats and kidney by ELISA inhibition assay. Cross-reactivity between αGal and pork kidney was studied with the ImmunoCAP(®) inhibition assay. RESULTS: Among the 14 patients, 12 presented with anaphylactic shock. Reactions occurred within 2 h from exposure in 67% of patients. Associated risk factors were observed in 10 cases, and alcohol was the main cofactor. Three patients underwent an oral challenge to pork kidney, and anaphylaxis occurred after ingestion of small quantities (1-2 g). Prick tests to kidney were positive in 54% of patients. All tested patients showed positive skin tests to Erbitux(®). All patients tested positive for IgE to αGal, with levels ranging from 0.4 to 294 kU/l. IgE binding to αGal was inhibited by raw pork kidney extract (mean, 77%; range, 55-87%), which showed a high amount of αGal determinants. CONCLUSIONS: Pork or beef kidney anaphylaxis is related to αGal IgE. Its peculiar severity could be due to an elevated content of αGal epitopes in kidney.

Benzo[a]pyrene (B[a]P) is a common food pollutant that causes DNA adduct formation and is carcinogenic. The report of a positive correlation between human plasma B[a]P levels and body mass index, together with B[a]P's lipophilicity, led us to test for possible adverse effects of B[a]P on adipose tissue. In ex vivo experiments using primary murine adipocytes, B[a]P rapidly (within minutes) and directly inhibited epinephrine-induced lipolysis (up to 75%) in a dose-dependent manner. Half-maximum inhibition was obtained with a B[a]P concentration of 0.9 mg.L(-1) (3.5 microm). Lipolysis induced by beta(1)-, beta(2)- and beta(3)-adrenoreceptor-specific agonists, as well as ACTH, were also significantly inhibited by B[a]P, whereas forskolin-induced lipolysis was not B[a]P-sensitive. Similar inhibition of catecholamine-induced lipolysis by B[a]P was also seen in isolated human adipocytes; half-maximum inhibition of lipolysis was achieved with a B[a]P concentration of 0.02 mg.L(-1) (0.08 microm). In vivo treatment of C57Bl/6J mice with 0.4 mg.kg(-1) B[a]P inhibited epinephrine-induced release of free fatty acids by 70%. Chronic exposure of mice to B[a]P (0.5 mg.kg(-1) injected i.p. every 48 h) for 15 days also decreased lipolytic response to epinephrine and induced a 43% higher weight gain compared with controls (B[a]P: 2.23 +/- 0.12 g versus control: 1.56 +/- 0.18 g, P < 0.01) due to increased fat mass. The weight gain occurred consistently without detectable changes in food intake. These results reveal a novel molecular mechanism of toxicity for the environmental pollutant B[a]P and introduce the notion that chronic exposure of human population to B[a]P and possibly other polycyclic aromatic hydrocarbons could have an impact on metabolic disorders, such as obesity.

Alterations in predation pressure can have large effects on trophically-structured systems. Modification of predator behaviour via ocean warming has been assessed by laboratory experimentation and metabolic theory. However, the influence of ocean acidification with ocean warming remains largely unexplored for mesopredators, including experimental assessments that incorporate key components of the assemblages in which animals naturally live. We employ a combination of long-term laboratory and mesocosm experiments containing natural prey and habitat to assess how warming and acidification affect the development, growth, and hunting behaviour in sharks. Although embryonic development was faster due to temperature, elevated temperature and CO2 had detrimental effects on sharks by not only increasing energetic demands, but also by decreasing metabolic efficiency and reducing their ability to locate food through olfaction. The combination of these effects led to considerable reductions in growth rates of sharks held in natural mesocosms with elevated CO2, either alone or in combination with higher temperature. Our results suggest a more complex reality for predators, where ocean acidification reduces their ability to effectively hunt and exert strong top-down control over food webs.

Furanocoumarins constitute a sub-family of coumarin compounds with important defense properties against pathogens and insects, as well as allelopathic functions in plants. Furanocoumarins are divided into two sub-groups according to the alignment of the furan ring with the lactone structure: linear psoralen and angular angelicin derivatives. Determination of furanocoumarin type is based on the prenylation position of the common precursor of all furanocoumarins, umbelliferone, at C6 or C8, which gives rise to the psoralen or angelicin derivatives, respectively. Here, we identified a membrane-bound prenyltransferase PcPT from parsley (Petroselinum crispum), and characterized the properties of the gene product. PcPT expression in various parsley tissues is increased by UV irradiation, with a concomitant increase in furanocoumarin production. This enzyme has strict substrate specificity towards umbelliferone and dimethylallyl diphosphate, and a strong preference for the C6 position of the prenylated product (demethylsuberosin), leading to linear furanocoumarins. The C8-prenylated derivative (osthenol) is also formed, but to a much lesser extent. The PcPT protein is targeted to the plastids in planta. Introduction of this PcPT into the coumarin-producing plant Ruta graveolens showed increased consumption of endogenous umbelliferone. Expression of PcPT and a 4-coumaroyl CoA 2'-hydroxylase gene in Nicotiana benthamiana, which does not produce furanocoumarins, resulted in formation of demethylsuberosin, indicating that furanocoumarin production may be reconstructed by a metabolic engineering approach. The results demonstrate that a single prenyltransferase, such as PcPT, opens the pathway to linear furanocoumarins in parsley, but may also catalyze the synthesis of osthenol, the first intermediate committed to the angular furanocoumarin pathway, in other plants.

Coumarins are important compounds that contribute to the adaptation of plants to biotic or abiotic stresses. Among coumarins, umbelliferone occupies a pivotal position in the plant phenylpropanoid network. Previous studies indicated that umbelliferone is derived from the ortho-hydroxylation of p-coumaric acid by an unknown biochemical step to yield 2,4-dihydroxycinnamic acid, which then undergoes spontaneous lactonization. Based on a recent report of a gene encoding a 2-oxoglutarate-dependent dioxygenase from Arabidopsis thaliana that exhibited feruloyl CoA 6'-hydroxylase activity (Bourgaud et al., 2006), we combined a bioinformatic approach and a cDNA library screen to identify an orthologous ORF (Genbank accession number JF799117) from Ruta graveolens L. This ORF shares 59% amino acid identity with feruloyl CoA 6'-hydroxylase, was functionally expressed in Escherichia coli, and converted feruloyl CoA into scopoletin and p-coumaroyl CoA into umbelliferone with equal activity. Its bi-functionality was further confirmed in planta: transient expression of JF799117 in Nicotiana benthamiana yielded plants with leaves containing high levels of umbelliferone and scopoletin when compared to control plants, which contained barely detectable traces of these compounds. The expression of JF799117 was also tightly correlated to the amount of umbelliferone that was found in UV-elicited R. graveolens leaves. Therefore, JF799117 encodes a p-coumaroyl CoA 2'-hydroxylase in R. graveolens, which represents a previously uncharacterized step in the synthesis of umbelliferone in plants. Psoralen, which is an important furanocoumarin in R. graveolens, was found to be a competitive inhibitor of the enzyme, and it may exert this effect through negative feedback on the enzyme at an upstream position in the pathway.

Beta-carotene acts as a pro-vitamin A or anti-cancer compound. Carrots contain the highest amount of beta-carotene of common fruits and vegetables, but each year 25% of carrot production is lost in the U.S. during processing and storage, while, at the same time, the market demand increases. This article is a review of the most recent studies concerning beta-carotene retention in carrots during processing and storage. Reducing the water activity by adding some aw lowering ingredients results in poor shelf-life. Drying or freezing gives better retention during storage than reducing the water activity, if the process is well controlled. Canning or freeze-drying were shown to be more effective. The trans form of beta-carotene in carrots is replaced by the cis form during processing. Beta-Carotene can be extracted from carrots, but the half-life of free beta-carotene is reduced to 2 d in the juice extract at room temperature. By encapsulation methods,the half-life can be increased by 6 months.

Abstract A database including results from 102 000 soil analyses was built for arable topsoil in France. We show a strong effect of soil particle‐size distribution on present organic carbon (OC) stocks in these soils. By using the upper decile values by textural classes as a proxy of the maximal carbon storage potential, we show that it might be more efficient to encourage practices favouring carbon accumulation in soils already having high OC stocks than to try to increase OC stocks in soils where present stocks are low.

Summary The mineral forms of phosphorus in three urban sewage sludges were characterized using high‐resolution solid‐state phosphorus‐31 nuclear magnetic resonance (NMR) coupled to a sequential extraction. The sludges studied were an anaerobically‐digested and heat‐treated sludge (Paris‐Achères), an activated sludge (Briare) and an anaerobically‐digested sludge (Nancy). NMR observations were conducted using both single‐pulse and cross‐polarization sequences in order to distinguish between 31 P nuclei far from 1 H nuclei, and 31 P located within a fraction of a nanometre of 1 H. This approach showed that a complex mixture of P species was present in these sludges. A mixture of hydrogenated octocalcium phosphates and apatites was observed in the three samples. Monetite was present in the anaerobically‐digested sludge and brushite in the activated sludge. Dehydrogenated condensed calcium phosphates (compounds with a Ca:P ratio higher than 1.0 such as fluorapatite or tricalcium phosphate) and dehydrogenated pyrophosphates were also probably present in the anaerobically‐digested sludge. A poorly‐ordered wavellite was observed in the three sludges after the HCl extraction. However, results were inconclusive as to whether this mineral was present in the three sludges, or had been precipitated during the sequential extraction.

OBJECTIVE: To compare late-preterm infants (33-36 weeks) with term infants (≥37 weeks) on incidence of morbidities in the first 3 years of life and healthcare costs during the first 2 years of life and third year of life. METHODS: Administrative health records of live infants born between January 1, 1997, and December 31, 2000 with 3 years follow-up data (N=35733) were linked. First, diagnoses of morbidities were compared between late-preterm and term infants using Cox's proportional hazards models. Healthcare costs expressed as mean total costs and cost ratios, accrued following initial hospital discharge after birth, were also examined. RESULTS: The three most common reasons for hospitalisation in late-preterm and term infants were acute bronchitis, otitis media and pneumonia. The most frequent reasons for physician visits included acute upper respiratory infections, otitis media and bronchiolitis. The highest HR were detected for chronic bronchitis 1.64 (1.13-2.39), hearing loss 1.56 (1.14-2.15) and bacterial diseases 1.28 (1.09-1.49). The mean total cost for late-preterm infants during the first 2 years of life was $2568 CAD compared with $1285 CAD for term infants, cost ratio =1.99 (95% CI 1.90 to 2.09). In the third year of life, the cost ratio reduced to 1.46 (95% CI 1.39 to 1.54). CONCLUSIONS: Late-preterm infants are at higher risk of specific morbidities compared with term infants. Their mean total costs fall from almost double that of term infants during the first 2 years of life, to just 46% greater in the third year of life.

Because soils are both a source and a sink for atmospheric CO2, there is an increasing need to characterize the spatial distribution of soil C pools. Large amounts of organic carbon (OC) accumulate in hydric bottom-lands soils. In the Armorican Massif (Western France) where these soils represent 20% of the total surface area, the spatial characterization of OC pools is difficult to assess due to methodological problems such as high spatial variability. Soil color indexes, which combine various characteristics of soil horizons or profiles, are an alternative approach for quantifying the differences in OC storage. In addition, terrain attributes derived from Digital Elevation Models (DEM) may be useful in characterizing the distribution of soil color indexes over large areas. Thus, the overall goal of this work was the development and application of a model for use in predicting the organic carbon (OC) content of soil areas. To accomplish this, extensive examination of soil morphology combined with selected terrain attributes measured in the field and calculated from a digital elevation model (DEM) were used. Soil samples were collected in Western France from a 2-ha agricultural parcel that forms the major part of a hillslope. The results indicate that OC stocks of the entire profile were correlated highly to a soil hydromorphic index (HI) (r2 = 0.80). HI is a function of the percent of the total soil profile depth constituted by horizons with some degree of hydromorphic feature development and the moist color of the surface A horizon. Using a stepwise regression technique, we constructed a prediction model of HI distribution by using the relations between HI and (i) the elevation above the stream bank (ES) (r2 = 0.80); (ii) the downslope gradient (DG) (r2 = 0.55); and (iii) the upslope contributing area (AMU) (r2 = 0.60). Validation of this model on a second site showed that topographical attributes explained up to 75% of the profile OC stock variability. These results confirmed that the integration of a soil index and topographical information is a useful tool for prediction of OC distribution. In addition, the use of soil morphologic indexes could significantly improved the construction and the validation of soil-landscape models because it would minimize laboratory measurements of OC reservoirs.

Abstract Sewage sludges have been used for many years as sources of P for agricultural crops, but there is a lack of information regarding the proportion of sludge P that can be used by crops. The aim of this work was to assess the importance of soil available P and sludge origin on the utilization of sludge P by plants. First, the changes in soil P isotopically exchangeable within 1 min (E1min) were measured in incubated soil‐sludge mixtures using two soils and four sludges. Then, the uptake of sludge P by ryegrass ( Lolium perenne L.) was measured on the same sludge‐amended soils. The application of sludges increased E1min to values lower or equal to those obtained following the application monocalcium phosphate. Similarly the utilization of sludge P by ryegrass was systematically lower than the utilization of P derived from a water‐soluble fertilizer. In both soils, the lowest utilization of sludge P was observed for the two FeSO 4 flocculated and anaerobically digested sludges, while the primary sludge and the aerobically digested sludge released somewhat higher quantities of P to ryegrass. In the clayey soil, the amount of sludge P taken up by the crop was significantly related both to the sludge and soil available P content, whereas no such relation was observed in the loamy soil because of its high available P content. The origin of the sludge and the soil available P content must therefore be taken into account when advising sludge application to crops to adjust P inputs to plant needs.

Abstract Microcapsules are produced by a dropwise addition of one solution into a solidifying bath. By applying an electrostatic potential between the droplet formation device and the collecting solution, it is possible to obtain smaller droplets which are desirable for many applications. Droplet formation may be divided into two phases. Under a certain critical value of the electric potential U c , liquid exits the nozzle as droplets. The surface tension decreases with increasing electric potential resulting in a reduction of droplet diameter to approximately 200 μm. At higher electric potential, liquid exits the nozzle as a jet which subsequently breaks into droplets, which are smaller than 200 μm. In this case, droplet size is mainly determined by the jet instability (theory of Rayleigh).

BACKGROUND: Attrition in cohort studies challenges causal inference. Although inverse probability weighting (IPW) has been proposed to handle attrition in association analyses, its relevance has been little studied in this context. We aimed to investigate its ability to correct for selection bias in exposure-outcome estimation by addressing an important methodological issue: the specification of the response model. METHODS: A simulation study compared the IPW method with complete-case analysis (CCA) for nine response-mechanism scenarios (3 missing at random - MAR and 6 missing not at random - MNAR). Eighteen response models differing by the type of variables included were assessed. RESULTS: The IPW method was equivalent to CCA in terms of bias and consistently less efficient in all scenarios, regardless of the response model tested. The most effective response model included only the confounding factors of the association model. CONCLUSION: Our study questions the ability of the IPW method to correct for selection bias in situations of attrition leading to missing outcomes. If the method is to be used, we encourage including only the confounding variables of the association of interest in the response model.

We present a convolutional neural network to classify distinct cosmological scenarios based on the statistically similar weak-lensing maps they generate. Modified gravity (MG) models that include massive neutrinos can mimic the standard concordance model [Lambda cold dark matter ($\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$)] in terms of Gaussian weak-lensing observables. An inability to distinguish viable models that are based on different physics potentially limits a deeper understanding of the fundamental nature of cosmic acceleration. For a fixed redshift of sources, we demonstrate that a machine learning network trained on simulated convergence maps can discriminate between such models better than conventional higher-order statistics. Results improve further when multiple source redshifts are combined. To accelerate training, we implement a novel data compression strategy that incorporates our prior knowledge of the morphology of typical convergence map features. Our method fully distinguishes $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$ from its most similar MG model on noise-free data, and it correctly identifies among the MG models with at least 80% accuracy when using the full redshift information. Adding noise lowers the correct classification rate of all models, but the neural network still significantly outperforms the peak statistics used in a previous analysis.

Plantaricin SA6, a bacteriocin produced by Lactobacillus plantarum SA6, exhibited an inhibitory action against several mesophilic lactobacilli. It was stable at 90-100 degrees C at pH 2-4 and it remained stable in the presence of several organic solvents, urea or beta-mercaptoethanol. Plantaricin SA6 bound specifically to the cell surface of only plantaricin SA6-sensitive bacteria. The putative receptors are not destroyed by different hydrolytic enzymes added to the phosphate buffer. Plantaricin SA6 acted as a bactericidal agent lysing sensitive strains, that became more permeable to ortho-nitro-phenol-beta-galactoside and lost their intracellular K+ ions and u.v.-absorbing materials. Both the adsorption and lethal action of plantaricin SA6 were maximal between pH 4 and 7, but the range of temperature tested (5-37 degrees C) had no effect. Ions (of several salts such as MgCl2) inhibited the binding of plantaricin SA6 and protected cells against bacteriocin action.

Background: This study investigated the periodization of elite swimmers’ training over the 25 weeks preceding the major competition of the season. Methods: We conducted a retrospective observational study of elite male (n = 60) and female (n = 67) swimmers (46 sprint, 81 middle-distance) over 20 competitive seasons (1992-2012). The following variables were monitored: training corresponding to blood lactate 4‒6 mmol.L-1, >6 mmol.L-1, and maximal swimming speed; general conditioning and maximal strength training hours; total training load (TTL); and the mean normalised volumes for both in-water and dryland workouts. Latent class mixed modelling was used to identify various TTL pattern groups. The associations between pattern groups and sex, age, competition event, Olympic quadrennial year, training contents, and relative performance were quantified. Results: For the entire cohort, ~86‒90% of the training was swum at an intensity of [La]b ≤4 mmol.L-1. This training volume was divided into 40‒44% at 4‒6 mmol.L-1 and 3.5‒4.5% at >6 mmol.L-1. Three sprint TTL patterns were identified: a pattern with two long ~14‒15-week macrocycles, one with two ~12‒13 week macrocycles each composed of a balanced training load, and one with a single stable flat macrocycle. The long pattern elicited the fastest performances and was most prevalent in Olympic quadrennials (i.e., 4 seasons preceding the 2004, 2008 and 2012 Olympic Games). This pattern exhibited moderate week-to-week TTL variability (6±3%), progressive training load increases between macrocycles, and more training at ≤4 mmol.L-1 and >6 mmol.L-1. This fastest sprint pattern showed a waveform in the second macrocycle consisting of two progressive load peaks 10‒11 and 4‒5‒6 weeks before competition. The stable flat pattern was the slowest and showed low TTL variability (4±3%), training load decreases between macrocycles (P6 mmol.L-1, were associated with peak performance in elite swimmers.