Sélection de modèle d'imputation à partir de modèles bayésiens hiérarchiques linéaires multivariés

Les logiciels utilisés sont Splus et R.

Imputation en présence de données contenant des zéros

L’imputation simple est très souvent utilisée dans les enquêtes pour compenser pour la non-réponse partielle. Dans certaines situations, la variable nécessitant l’imputation prend des valeurs nulles un très grand nombre de fois. Ceci est très fréquent dans les enquêtes entreprises qui collectent les variables économiques. Dans ce mémoire, nous étudions les propriétés de deux méthodes d’imputation souvent utilisées en pratique et nous montrons qu’elles produisent des estimateurs imputés biaisés en général. Motivé par un modèle de mélange, nous proposons trois méthodes d’imputation et étudions leurs propriétés en termes de biais. Pour ces méthodes d’imputation, nous considérons un estimateur jackknife de la variance convergent vers la vraie variance, sous l’hypothèse que la fraction de sondage est négligeable. Finalement, nous effectuons une étude par simulation pour étudier la performance des estimateurs ponctuels et de variance en termes de biais et d’erreur quadratique moyenne.

When zero doesn't mean it and other geomathematical mischief

There is almost not a case in exploration geology, where the studied data doesn’t includes below detection limits and/or zero values, and since most of the geological data responds to lognormal distributions, these “zero data” represent a mathematical challenge for the interpretation. We need to start by recognizing that there are zero values in geology. For example the amount of quartz in a foyaite (nepheline syenite) is zero, since quartz cannot co-exists with nepheline. Another common essential zero is a North azimuth, however we can always change that zero for the value of 360°. These are known as “Essential zeros”, but what can we do with “Rounded zeros” that are the result of below the detection limit of the equipment? Amalgamation, e.g. adding Na2O and K2O, as total alkalis is a solution, but sometimes we need to differentiate between a sodic and a potassic alteration. Pre-classification into groups requires a good knowledge of the distribution of the data and the geochemical characteristics of the groups which is not always available. Considering the zero values equal to the limit of detection of the used equipment will generate spurious distributions, especially in ternary diagrams. Same situation will occur if we replace the zero values by a small amount using non-parametric or parametric techniques (imputation). The method that we are proposing takes into consideration the well known relationships between some elements. For example, in copper porphyry deposits, there is always a good direct correlation between the copper values and the molybdenum ones, but while copper will always be above the limit of detection, many of the molybdenum values will be “rounded zeros”. So, we will take the lower quartile of the real molybdenum values and establish a regression equation with copper, and then we will estimate the “rounded” zero values of molybdenum by their corresponding copper values. The method could be applied to any type of data, provided we establish first their correlation dependency. One of the main advantages of this method is that we do not obtain a fixed value for the “rounded zeros”, but one that depends on the value of the other variable. Key words: compositional data analysis, treatment of zeros, essential zeros, rounded zeros, correlation dependency

Adjusted jackknife for imputation under unequal probability sampling without replacement

Imputation is commonly used to compensate for item non-response in sample surveys. If we treat the imputed values as if they are true values, and then compute the variance estimates by using standard methods, such as the jackknife, we can seriously underestimate the true variances. We propose a modified jackknife variance estimator which is defined for any without-replacement unequal probability sampling design in the presence of imputation and non-negligible sampling fraction. Mean, ratio and random-imputation methods will be considered. The practical advantage of the method proposed is its breadth of applicability.

EDITOR'S CHOICE: REVIEW: Trait matching of flower visitors and crops predicts fruit set better than trait diversity

Understanding the relationships between trait diversity, species diversity and ecosystem functioning is essential for sustainable management. For functions comprising two trophic levels, trait matching between interacting partners should also drive functioning. However, the predictive ability of trait diversity and matching is unclear for most functions, particularly for crop pollination, where interacting partners did not necessarily co-evolve. World-wide, we collected data on traits of flower visitors and crops, visitation rates to crop flowers per insect species and fruit set in 469 fields of 33 crop systems. Through hierarchical mixed-effects models, we tested whether flower visitor trait diversity and/or trait matching between flower visitors and crops improve the prediction of crop fruit set (functioning) beyond flower visitor species diversity and abundance. Flower visitor trait diversity was positively related to fruit set, but surprisingly did not explain more variation than flower visitor species diversity. The best prediction of fruit set was obtained by matching traits of flower visitors (body size and mouthpart length) and crops (nectar accessibility of flowers) in addition to flower visitor abundance, species richness and species evenness. Fruit set increased with species richness, and more so in assemblages with high evenness, indicating that additional species of flower visitors contribute more to crop pollination when species abundances are similar. Synthesis and applications. Despite contrasting floral traits for crops world-wide, only the abundance of a few pollinator species is commonly managed for greater yield. Our results suggest that the identification and enhancement of pollinator species with traits matching those of the focal crop, as well as the enhancement of pollinator richness and evenness, will increase crop yield beyond current practices. Furthermore, we show that field practitioners can predict and manage agroecosystems for pollination services based on knowledge of just a few traits that are known for a wide range of flower visitor species.

Analysis of the electrical conductivity in milking fractions as a mean for detecting and characterizing mastitis in goats

The aim of the work was to study the effect of milking fraction on electrical conductivity of milk (EC) to improve its use in dairy goat mastitis detection using automatic EC measurements during milking. The experiment was carried out on a group of 84 Murciano-Granadina goats (28 primiparous and 56 multiparous). Goats were in the fourth month of lactation. A linear mixed model was used to analyse the relationship between EC or somatic cell count (SCC) of gland milk and parity, mammary gland health status, analysed fraction (first 100 mL=F-1; machine milk=F-2; and stripping milk=F-3) and their first order interactions. Additionally, the mastitis detection characteristics (sensitivity, specificity, positive predictive value and negative predictive value) of SCC and EC were studied at different thresholds.All factors considered were significant for EC and SCC. EC decreased significantly as milking progressed (from F-1 to F-3) in both healthy and infected glands. EC was not significantly different between healthy and infected glands in F-1 and F-2 fractions, but EC of healthy glands (5.01 mS/cm) was significantly lower than in infected glands (5.03 mS/cm) at F-3.Mastitis detection characteristics of EC did not differ amongst studied fractions. The small significant difference of EC between healthy and infected glands obtained in F-3 fraction did not yield better sensitivity results compared to F-1 and F-2. The best EC mastitis detection characteristics were obtained at 5.20 mS/cm threshold (sensitivity of 70% and specificity of 50%). The best SCC mastitis detection characteristics were obtained at 300,000 cells/mL threshold and F-3 fraction (sensitivity of 85% and specificity of 65%).It was concluded that mastitis detection characteristics of EC were similar in the three milking fractions analysed, being slightly better for SCC in F-3 fraction. As shown in previous studies, there are no factors other than the mammary gland health status that affect milk EC and should be considered in the algorithms for mastitis detection to improve the results. (C) 2012 Elsevier B.V. All rights reserved.

Sensitivity, specificity and predictive values of linear and nonlinear indices of heart rate variability in stable angina patients

Background: Decreased heart rate variability (HRV) is related to higher morbidity and mortality. In this study we evaluated the linear and nonlinear indices of the HRV in stable angina patients submitted to coronary angiography. Methods. We studied 77 unselected patients for elective coronary angiography, which were divided into two groups: coronary artery disease (CAD) and non-CAD groups. For analysis of HRV indices, HRV was recorded beat by beat with the volunteers in the supine position for 40 minutes. We analyzed the linear indices in the time (SDNN [standard deviation of normal to normal], NN50 [total number of adjacent RR intervals with a difference of duration greater than 50ms] and RMSSD [root-mean square of differences]) and frequency domains ultra-low frequency (ULF) ≤ 0,003 Hz, very low frequency (VLF) 0,003 - 0,04 Hz, low frequency (LF) (0.04-0.15 Hz), and high frequency (HF) (0.15-0.40 Hz) as well as the ratio between LF and HF components (LF/HF). In relation to the nonlinear indices we evaluated SD1, SD2, SD1/SD2, approximate entropy (-ApEn), α1, α2, Lyapunov Exponent, Hurst Exponent, autocorrelation and dimension correlation. The definition of the cutoff point of the variables for predictive tests was obtained by the Receiver Operating Characteristic curve (ROC). The area under the ROC curve was calculated by the extended trapezoidal rule, assuming as relevant areas under the curve ≥ 0.650. Results: Coronary arterial disease patients presented reduced values of SDNN, RMSSD, NN50, HF, SD1, SD2 and -ApEn. HF ≤ 66 ms§ssup§2§esup§, RMSSD ≤ 23.9 ms, ApEn ≤-0.296 and NN50 ≤ 16 presented the best discriminatory power for the presence of significant coronary obstruction. Conclusion: We suggest the use of Heart Rate Variability Analysis in linear and nonlinear domains, for prognostic purposes in patients with stable angina pectoris, in view of their overall impairment. © 2012 Pivatelli et al.; licensee BioMed Central Ltd.

Predictive formulas for food base excess and urine pH estimations of cats

Food base excess (BE, mEq/kg) can be calculated from the diet macroelements, together with either the sulfur amino acids methionine and cysteine (BEaa) or total sulfur (BEs) concentrations. The present study compared the use of sulfur or methionine and cysteine for calculating the food BE (experiment 1) and investigated the influence of food BE on blood gas analysis and the urine pH of cats, and proposes a prediction equation to estimate the urine pH of cats fed kibble diets based on the calculated food BE (experiments 2 and 3). In experiment 1, nine healthy, adult cats were used in a change-over design and fed with nine commercial dry cat foods. The cats were housed in metabolism cages over seven days for adaptation and three days for total urine collection. All of the urine produced over 24h was pooled by cat and diet. The cats' acid-base status was assessed through blood gas analysis after 10 days of diet consumption. A mean difference of -115mEq/kg between BEs and BEaa was observed, which could be explained by a greater concentration of sulfur in the whole diet than in methionine and cysteine. Urine pH presented a stronger correlation with food BEs (R2=0.95; P<0.001) than with food BEaa (R2=0.86; P<0.001). Experiment 2 included 30 kibble diets, and each diet was tested in six cats. The food BEs varied between -180 and +307mEq/kg, and the urine pH of the cats varied between 5.60 and 7.74. A significant correlation was found between the measured urine pH and the food BEs (urinary pH=6.269+[0.0036×BEs]+[0.000003×BEs2]; R2=0.91; P<0.001). In experiment 3, eight kibble diets were tested (food BEs between -187mEq/kg and +381mEq/kg) to validate the equation proposed in experiment 2 and to compare the obtained results with previously published formulae. The results of the proposed formula presented a high concordance correlation coefficient (0.942) and high accuracy (0.979) with the measured values, and the estimates of urine pH did not differ from the values obtained in cats (P>0.05). The cats' venous blood pH, bicarbonate, and blood BE were correlated with food BEs (P<0.001); the consumption of diets with low food BEs induced a reduction in these parameters. In conclusion, food macroelement composition has a strong influence on cats' acid-base equilibrium and food BEs calculation is a useful tool to formulate and balance kibble diets for felines. © 2013 Elsevier B.V.

Mean surface temperature prediction models for broiler chickens-a study of sensible heat flow

Body surface temperature can be used to evaluate thermal equilibrium in animals. The bodies of broiler chickens, like those of all birds, are partially covered by feathers. Thus, the heat flow at the boundary layer between broilers' bodies and the environment differs between feathered and featherless areas. The aim of this investigation was to use linear regression models incorporating environmental parameters and age to predict the surface temperatures of the feathered and featherless areas of broiler chickens. The trial was conducted in a climate chamber, and 576 broilers were distributed in two groups. In the first trial, 288 broilers were monitored after exposure to comfortable or stressful conditions during a 6-week rearing period. Another 288 broilers were measured under the same conditions to test the predictive power of the models. Sensible heat flow was calculated, and for the regions covered by feathers, sensible heat flow was predicted based on the estimated surface temperatures. The surface temperatures of the feathered and featherless areas can be predicted based on air, black globe or operative temperatures. According to the sensible heat flow model, the broilers' ability to maintain thermal equilibrium by convection and radiation decreased during the rearing period. Sensible heat flow estimated based on estimated surface temperatures can be used to predict animal responses to comfortable and stressful conditions. © 2013 ISB.

Comparison of predictive equations for resting energy expenditure in overweight and obese adults

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Sensitivity, specificity and predictive values of linear and nonlinear indices of heart rate variability in stable angina patients

Abstract Background Decreased heart rate variability (HRV) is related to higher morbidity and mortality. In this study we evaluated the linear and nonlinear indices of the HRV in stable angina patients submitted to coronary angiography. Methods We studied 77 unselected patients for elective coronary angiography, which were divided into two groups: coronary artery disease (CAD) and non-CAD groups. For analysis of HRV indices, HRV was recorded beat by beat with the volunteers in the supine position for 40 minutes. We analyzed the linear indices in the time (SDNN [standard deviation of normal to normal], NN50 [total number of adjacent RR intervals with a difference of duration greater than 50ms] and RMSSD [root-mean square of differences]) and frequency domains ultra-low frequency (ULF) ≤ 0,003 Hz, very low frequency (VLF) 0,003 – 0,04 Hz, low frequency (LF) (0.04–0.15 Hz), and high frequency (HF) (0.15–0.40 Hz) as well as the ratio between LF and HF components (LF/HF). In relation to the nonlinear indices we evaluated SD1, SD2, SD1/SD2, approximate entropy (−ApEn), α1, α2, Lyapunov Exponent, Hurst Exponent, autocorrelation and dimension correlation. The definition of the cutoff point of the variables for predictive tests was obtained by the Receiver Operating Characteristic curve (ROC). The area under the ROC curve was calculated by the extended trapezoidal rule, assuming as relevant areas under the curve ≥ 0.650. Results Coronary arterial disease patients presented reduced values of SDNN, RMSSD, NN50, HF, SD1, SD2 and -ApEn. HF ≤ 66 ms2, RMSSD ≤ 23.9 ms, ApEn ≤−0.296 and NN50 ≤ 16 presented the best discriminatory power for the presence of significant coronary obstruction. Conclusion We suggest the use of Heart Rate Variability Analysis in linear and nonlinear domains, for prognostic purposes in patients with stable angina pectoris, in view of their overall impairment.

Immunohistochemical and Molecular Prognostic/Predictive Markers in Neoplastic Diseases

Traditional morphological examinations are not anymore sufficient for a complete evaluation of tumoral tissue and the use of neoplastic markers is of utmost importance. Neoplastic markers can be classified in: diagnostic, prognostic and predictive markers. Three markers were analyzed. 1) Insulin-like growth factor binding protein 2 (IGFBP2) was immunohistochemically examined in prostatic tissues: 40 radical prostatectomies from hormonally untreated patients with their preoperative biopsies, 10 radical prostatectomies from patients under complete androgen ablation before surgery and 10 simple prostatectomies from patients with bladder outlet obstruction. Results were compared with α-methylacyl-CoA racemase (AMACR). IGFBP2 was expressed in the cytoplasm of untreated adenocarcinomas and, to a lesser extent, in HG-PIN; the expression was markedly lower in patients after complete androgen ablation. AMACR was similarly expressed in both adenocarcinoma and HG-PIN, the level being similar in both lesions; the expression was slightly lower in patients after complete androgen ablation. IGFBP2 may be used a diagnostic marker of prostatic adenocarcinomas. 2) Heparan surface proteoglycan immunohistochemical expression was examined in 150 oral squamous cell carcinomas. Follow up information was available in 93 patients (range: 6-34 months, mean: 19±7). After surgery, chemotherapy was performed in 8 patients and radiotherapy in 61 patients. Multivariate and univariate overall survival analyses showed that high expression of syndecan-1 (SYN-1) was associated with a poor prognosis. In patients treated with radiotherapy, such association was higher. SYN-1 is a prognostic marker in oral squamous cell carcinomas; it may also represent a predictive factor for responsiveness to radiotherapy. 3) EGFR was studied in 33 pulmonary adenocarcinomas with traditional DNA sequencing methods and with two mutation-specific antibodies. Overall, the two antibodies had 61.1% sensitivity and 100% specificity in detecting EGFR mutations. EGFR mutation-specific antibodies may represent a predictive marker to identify patients candidate to tyrosine kinase inhibitors therapy.

The vulnerable carotid plaque. Identification of endothelial markers predictive of plaque weakness, rupture predisposition and neoangiogenesis

Background. Neoangiogenesis is crucial in plaque progression and instability. Previous data from our group demonstrated that intra-plaque neovessels show both a Nestin+/WT+ and a Nestin+/WT1- phenotype, the latter being correlated with complications and plaque instability. Aims. The aims of the present thesis are: (i) to confirm our previous results on Nestin/WT1 phenotype in a larger series of carotid atheromatous plaques, (ii) to evaluate the relationship between the Nestin+/WT1- neoangiogenesis phenotype and plaque morphology, (iii) to evaluate the relationship between the immunohistochemical and histopathological characteristics and the clinical instability of the plaques. Materials and Methods. Seventy-three patients (53 males, 20 females, mean age 71 years) were consecutively enrolled. Symptoms, brain CT scan, 14 histological variables, including intraplaque hemorrhage and diffuse calcifications, were collected. Immunohistochemistry for CD34, Nestin and WT1 was performed. RT-PCR was performed to evaluate Nestin and WT1 mRNA (including 5 healthy arteries as controls). Results. Diffusely calcified plaques (13 out of 73) were found predominantly in females (P=0.017), with a significantly lower incidence of symptoms (TIA/stroke) and brain focal lesions (P=0.019 and P=0.013 respectively) than not-calcified plaques, but with the same incidence of intraplaque complications (P=0.156). Accordingly, both calcified and not calcified plaques showed similar mean densities of positivity for CD34, Nestin and WT1. The density of Nestin and WT1 correlated with the occurrence of intra-plaque hemorrhage in all cases, while the density of CD34 correlated only in not-calcified plaques. Conclusions. We confirmed that the Nestin+/WT1- phenotype characterizes the neovessels of instable plaques, regardless the real amount of CD34-positive neoangiogenesis. The calcified plaques show the same incidence of histological complications, albeit they do not influence symptomatology and plaque vulnerability. Female patients show a much higher incidence of not-complicated or calcified plaques, receiving de facto a sort of protection compared to male patients.

Kernel density correlation based non-rigid point set matching for statistical model-based 2D/3D reconstruction

This paper presents a kernel density correlation based nonrigid point set matching method and shows its application in statistical model based 2D/3D reconstruction of a scaled, patient-specific model from an un-calibrated x-ray radiograph. In this method, both the reference point set and the floating point set are first represented using kernel density estimates. A correlation measure between these two kernel density estimates is then optimized to find a displacement field such that the floating point set is moved to the reference point set. Regularizations based on the overall deformation energy and the motion smoothness energy are used to constraint the displacement field for a robust point set matching. Incorporating this non-rigid point set matching method into a statistical model based 2D/3D reconstruction framework, we can reconstruct a scaled, patient-specific model from noisy edge points that are extracted directly from the x-ray radiograph by an edge detector. Our experiment conducted on datasets of two patients and six cadavers demonstrates a mean reconstruction error of 1.9 mm

Influence of deep sternal wound infection on long-term survival after cardiac surgery

BACKGROUND: This study aimed to investigate the influence of deep sternal wound infection on long-term survival following cardiac surgery. MATERIAL AND METHODS: In our institutional database we retrospectively evaluated medical records of 4732 adult patients who received open-heart surgery from January 1995 through December 2005. The predictive factors for DSWI were determined using logistic regression analysis. Then, each patient with deep sternal wound infection (DSWI) was matched with 2 controls without DSWI, according to the risk factors identified previously. After checking balance resulting from matching, short-term mortality was compared between groups using a paired test, and long-term survival was compared using Kaplan-Meier analysis and a Cox proportional hazard model. RESULTS: Overall, 4732 records were analyzed. The mean age of the investigated population was 69.3±12.8 years. DSWI occurred in 74 (1.56%) patients. Significant independent predictive factors for deep sternal infections were active smoking (OR 2.19, CI95 1.35-3.53, p=0.001), obesity (OR 1.96, CI95 1.20-3.21, p=0.007), and insulin-dependent diabetes mellitus (OR 2.09, CI95 1.05-10.06, p=0.016). Mean follow-up in the matched set was 125 months, IQR 99-162. After matching, in-hospital mortality was higher in the DSWI group (8.1% vs. 2.7% p=0.03), but DSWI was not an independent predictor of long-term survival (adjusted HR 1.5, CI95 0.7-3.2, p=0.33). CONCLUSIONS: The results presented in this report clearly show that post-sternotomy deep wound infection does not influence long-term survival in an adult general cardio-surgical patient population.