4^th International Conference and Exhibition on Biometrics & Biostatistics November 16-18, 2015 San Antonio, USA

Marco Giordan has completed his PhD in Statistics in 2007 from Padua University and then worked as Biostatistician in different research institutes. Actually, he is a Researcher in the group of Biostatistics and Data Management at Edmund Mach Foundation, an institute promoting research in Agriculture.

Biological data are very often produced in different non-comparable batches. For data with repeated measurements and for longitudinal data, the correlated nature of the samples must also be considered in the procedure for the removal of the batch effects. Current literature on the removal of batch effects, however, is mainly concerned with the analysis of experiments having an independent sampling of the subjects. We have developed a procedure based on a linear mixed model to remove the batch effects from correlated data. Our procedure provides a filtered data set that can be used for further analyses.

Vanda M Lourenço has completed her PhD in December 2011 from Instituto Superior Técnico of the University of Lisbon. She is an Assistant Professor at the Department of Mathematics of the Faculty of Sciences and Technology of the NOVA University of Lisbon. Her research interests include, among others, robust statistical methods in genetic studies, either concerning plant, animal or human genetics applications. She is, currently, also a PI of a project directed to plant applications, funded by the Portuguese Foundation for Science and Technology.

Genomic Prediction (GP) is used in plant breeding to help find the best genotypes for selection. Here, one of the most important measures of GP, whose estimation is of the utmost importance for Genomic Selection (GS), is predictive accuracy. Additionally, the models of choice for analyzing field data are regression models. However, models that lean on classical likelihood are known to perform poorly when their underlying assumptions are violated. For example, the violation of the normality assumption may lead to biased parameter estimates, derail the coefficient of determination, hinder accurate variable selection and, in the case of the linear mixed model, interfere with variance components estimation. In plant genetics, these biases often translate into inaccurate heritability and predictive ability estimates that may, in turn, negatively impact predictive accuracy and thus GS. Since phenotypic data are prone to contamination and, therefore, almost never confirm to the normality assumption, one way of improving GS necessarily involves the improvement of the methods used in the estimation of these aforementioned measures. Here, robust statistical methods find a natural framework, since they are known to overcome some of the handicaps of the classical methodology, one of which is the departure from the normality premise. Therefore, a robust approach to a recently proposed two-stage method from the literature is presented. The introduced robust methodology is compared with its classical counterpart through simulation. Finally, an example of application involving a maize data set is given so that the adequacy and usefulness of the robust approach may be validated.

Tsitsiashvili G Sh has completed his Doctorate thesis on physical and mathematical sciences. He is the Head of IAM FEB RAS laboratory on probability methods and systems analysis and has published more than 100 papers in reputed journals. The main sphere of his researches is applied probability models: queuing, reliability, insurance, epidemic, etc.

In this paper, we consider oriented graph with finite sets of nodes and edges. Such graphs appear in a description of protein networks. These graphs may be factorized by the cyclic equivalence and the partial order between equivalence classes may be introduced. A direct way to realize this procedure is the Boolean multiplication of the graph contiguity matrix, but such approach is too complicated. So, we construct a sequential algorithm of an oriented graph factorization with a square number of arithmetical operations by a number of graph nodes. A decrease of the calculation complexity is connected with an introduction of a partial order matrix which is defined recursively by assignment operations. To test suggested factorization algorithm, we consider the protein network Arabidopsis with 2824 nodes and 7500 edges. Using sufficiently powerful server, it is possible to realize the factorization procedure by the calculation of the matrixes during 21 days. Using the sequential algorithm, it is possible to fulfill the factorization procedure on a notebook during one and half hours. An analysis of a distribution of equivalence classes by numbers of their nodes shows that there is the equivalence class (a kernel) with 958 nodes which composes approximately 34 percent of all network vertices.

Nassima Kihal is a PhD student at the University Beb Ezzouar of Algeria and Research Intern at the University of Montreal. She has a particular interest in New Biometric data and multimodal biometric system for person recognition.\r\n

Biometrics offers a natural and reliable solution to certain aspects of identity management with fully automated or semi-automated schemes to recognize individuals based on their inherent physical and/or behavioral characteristics including fingerprint, face, voice, iris, signature, hand and ear. Our goal is to use cornea as a new biometric trait for identity authentication by modeling it in 3D geometry. For this reason, we propose to start with the realization of a new corneal biometric database. The acquisition of the cornea is carried out by a Pentacam Topographer (Oculus). The Pentacam measurement process takes less than two seconds and minute eye movements are captured and corrected simultaneously. By measuring 25,000 true elevation points, precise representation, repeatability and analysis are guaranteed. These data points are then used to generate corneal maps used for diagnosis and treatment. Our database contains 344 corneal topographies captured from 43 different people of different ages using a both eyes. For each eye, we captured two sessions of corneal topography. The time interval between the two sessions was equal or greater than one month. In each session, 8 corneal topographies (4 left eyes and 4 right eyes) were captured. These multiple measurements are useful to assess repeatability, performance, collectability, etc. The corneal shape was recorded as a uniformly spaced (X-Y) grid (image) of raw elevations (Z). This elevation can be represented with an appropriate mathematical model such as a Zernike polynomial expansion. The model parameters (e.g., Zernike coefficients) can then be post-processed (e.g., with linear discriminant analysis LDA)) to become a ‘signature’ or ‘template’ of the cornea. Our proposed database was tested for person recognition in a preliminary study and gave an impressive equal error rate of less than 1%.

N/A

In matched case-crossover studies, it is generally accepted that the covariates on which a case and associated controls are matched cannot exert a confounding effect on independent predictors included in the conditional logistic regression model. The conditional logistic regression model is not able to detect any effects associated with the matching covariates by stratum, such as time and spatial location. We propose an approach which allows us to simultaneously evaluate the following three features:\r\n1) Detect the parametric relationship between the predictor and binary outcomes,\r\n2) Evaluate semi-parametric relationships between the predictor and time, and\r\n3) Determine whether there is an effect modification due to spatial location for a reduced number of locations.\r\nWe demonstrate the accuracy of the estimation using simulation studies and an epidemiological example of a 1-4 bi-directional case-crossover study of childhood aseptic meningitis with drinking water turbidity.\r\n