Arthur Tenenhaus

Significant correlation between a set of genetic polymorphisms and a functional brain network revealed by feature selection and sparse Partial Least Squares

Brain imaging is increasingly recognised as an intermediate phenotype to understand the complex path between genetics and behavioural or clinical phenotypes. In this context, a first goal is to propose methods to identify the part of genetic variability that explains some neuroimaging variability. Classical univariate approaches often ignore the potential joint effects that may exist between genes or the potential covariations between brain regions. In this paper, we propose instead to investigate an exploratory multivariate method in order to identify a set of Single Nucleotide Polymorphisms (SNPs) covarying with a set of neuroimaging phenotypes derived from functional Magnetic Resonance Imaging (fMRI). Recently, Partial Least Squares (PLS) regression or Canonical Correlation Analysis (CCA) have been proposed to analyse DNA and transcriptomics. Here, we propose to transpose this idea to the DNA vs. imaging context. However, in very high-dimensional settings like in imaging genetics studies, such multivariate methods may encounter overfitting issues. Thus we investigate the use of different strategies of regularisation and dimension reduction techniques combined with PLS or CCA to face the very high dimensionality of imaging genetics studies. We propose a comparison study of the different strategies on a simulated dataset first and then on a real dataset composed of 94 subjects, around 600,000 SNPs and 34 functional MRI lateralisation indexes computed from reading and speech comprehension contrast maps. We estimate the generalisability of the multivariate association with a cross-validation scheme and demonstrate the significance of this link, using a permutation procedure. Univariate selection appears to be necessary to reduce the dimensionality. However, the significant association uncovered by this two-step approach combining univariate filtering and L1-regularised PLS suggests that discovering meaningful genetic associations calls for a multivariate approach.

Variable selection for generalized canonical correlation analysis

Regularized generalized canonical correlation analysis (RGCCA) is a generalization of regularized canonical correlation analysis to 3 or more sets of variables. RGCCA is a component-based approach which aims to study the relationships between several sets of variables. The quality and interpretability of the RGCCA components are likely to be affected by the usefulness and relevance of the variables in each block. Therefore, it is an important issue to identify within each block which subsets of significant variables are active in the relationships between blocks. In this paper, RGCCA is extended to address the issue of variable selection. Specifically, sparse generalized canonical correlation analysis (SGCCA) is proposed to combine RGCCA with an [Formula: see text]-penalty in a unified framework. Within this framework, blocks are not necessarily fully connected, which makes SGCCA a flexible method for analyzing a wide variety of practical problems. Finally, the versatility and usefulness of SGCCA are illustrated on a simulated dataset and on a 3-block dataset which combine gene expression, comparative genomic hybridization, and a qualitative phenotype measured on a set of 53 children with glioma. SGCCA is available on CRAN as part of the RGCCA package.

Gene Association Networks from Microarray Data Using a Regularized Estimation of Partial Correlation Based on PLS Regression

HASH(0x2980028)

Detection of melanoma from dermoscopic images of naevi acquired under uncontrolled conditions

Background and objective: Several systems for the diagnosis of melanoma from images of naevi obtained under controlled conditions have demonstrated comparable efficiency with dermatologists. However, their robustness to analyze daily routine images was sometimes questionable. The purpose of this work is to investigate to what extent the automatic melanoma diagnosis may be achieved from the analysis of uncontrolled images of pigmented skin lesions.

Regularized generalized canonical correlation analysis for multiblock or multigroup data analysis

This paper presents an overview of methods for the analysis of data structured in blocks of variables or in groups of individuals. More specifically, regularized generalized canonical correlation analysis (RGCCA), which is a unifying approach for multiblock data analysis, is extended to be also a unifying tool for multigroup data analysis. The versatility and usefulness of our approach is illustrated on two real datasets.

The Intracellular Localization of ID2 Expression Has a Predictive Value in Non Small Cell Lung Cancer

Background ID2 is a member of a subclass of transcription regulators belonging to the general bHLH (basic-helix-loop-helix) family of transcription factors. In normal cells, ID2 is responsible for regulating the balance between proliferation and differentiation. More recent studies have demonstrated that ID2 is involved in tumor progression in several cancer types such as prostate or breast. Methodology/Principal Findings In this work, we investigated, for the first time, the relationship between the expression of ID2 in non-small cell lung cancer (NSCLC) patients and the clinicopathological features and prognosis of these patients. Immunohistochemistry was performed on tissue microarrays, which included 62 NSCLC tumors. In malignant tissues, ID2 expression has been detected in both the nuclear and cytoplasmic compartments, but we have demonstrated that only nuclear expression of ID2 is inversely correlated with the differentiation grade of the tumor (p = 0.007). Interestingly, among patients with poorly differentiated tumors, high nuclear expression of ID2 was an independent and unfavorable prognostic factor for survival (p = 0.036). Conclusions These results suggest that ID2 could be involved in tumor dedifferentiation processes of NSCLC, and could be used as prognostic marker for patients with poorly differentiated tumors.

Kernel Generalized Canonical Correlation Analysis

There is a growing need to analyze datasets characterized by several sets of variables observed on a single set of observations. Such complex but structured dataset are known as multiblock dataset, and their analysis requires the development of new and flexible tools. For this purpose, Kernel Generalized Canonical Correlation Analysis (KGCCA) is proposed and offers a general framework for multiblock data analysis taking into account an a priori graph of connections between blocks. It appears that KGCCA subsumes, with a single monotonically convergent algorithm, a remarkably large number of well-known and new methods as particular cases. KGCCA is applied to a simulated 3 -block dataset and a real molecular biology dataset that combines Gene Expression data, Comparative Genomic Hybridization data and a qualitative phenotype measured for a set of 53 children with glioma.KGCCA is available on CRAN as part of the RGCCA package.

Compressed sensing based 3d tomographic reconstruction for rotational angiography

In this paper, we address three-dimensional tomographic reconstruction of rotational angiography acquisitions. In clinical routine, angular subsampling commonly occurs, due to the technical limitations of C-arm systems or possible improper injection. Standard methods such as filtered backprojection yield a reconstruction that is deteriorated by sampling artifacts, which potentially hampers medical interpretation. Recent developments of compressed sensing have demonstrated that it is possible to significantly improve reconstruction of subsampled datasets by generating sparse approximations through l1-penalized minimization. Based on these results, we present an extension of the iterative filtered backprojection that includes a sparsity constraint called soft background subtraction. This approach is shown to provide sampling artifact reduction when reconstructing sparse objects, and more interestingly, when reconstructing sparse objects over a non-sparse background. The relevance of our approach is evaluated in cone-beam geometry on real clinical data.

Imaging Genetics: Bio-Informatics and Bio-Statistics Challenges

The IMAGEN study—a very large European Research Project—seeks to identify and characterize biological and environmental factors that influence teenagers mental health. To this aim, the consortium plans to collect data for more than 2000 subjects at 8 neuroimaging centres. These data comprise neuroimaging data, behavioral tests (for up to 5 hours of testing), and also white blood samples which are collected and processed to obtain 650 k single nucleotide polymorphisms (SNP) per subject. Data for more than 1000 subjects have already been collected. We describe the statistical aspects of these data and the challenges, such as the multiple comparison problem, created by such a large imaging genetics study (i.e., 650 k for the SNP, 50 k data per neuroimage).We also suggest possible strategies, and present some first investigations using uni or multi-variate methods in association with re-sampling techniques. Specifically, because the number of variables is very high, we first reduce the data size and then use multivariate (CCA, PLS) techniques in association with re-sampling techniques.

Classification of segmental wall motion in echocardiography using quantified parametric images

The interpretation of cine-loops and parametric images to assess regional wall motion in echocardiography requires to acquire an expertise, which is based on training. To overcome the training phase for the interpretation of new parametric images, a quantification based on profiles in the parametric images was attempted. The classification of motion was performed on a training set including 362 segments and tested on a second database including 238 segments. The consensual visual interpretation of two-dimensional sequences by two experienced readers were used as the ”gold standard”. Mono- and multi-parametric classification approaches were undertaken. Results show an accuracy of 74% for training and 68% for test in case of mono-parametric approach. They are 80% and 67% in case of multi-parametric approach. Moreover, the evaluation protocol enables to understand the limitations of this approach. The in-depth study shows that a large part of false-positive segments are apical segments. This suggests that taking into account the segment location could improve the performances.