Gérard Ramstein

MADGene: retrieval and processing of gene identifier lists for the analysis of heterogeneous microarray datasets

Summary: MADGene is a software environment comprising a web-based database and a java application. This platform aims at unifying gene identifiers (ids) and performing gene set analysis. MADGene allows the user to perform inter-conversion of clone and gene ids over a large range of nomenclatures relative to 17 species. We propose a set of 23 functions to facilitate the analysis of gene sets and we give two microarray applications to show how MADGene can be used to conduct meta-analyses. Availability: The MADGene resources are freely available online from http://www.madtools.org, a website dedicated to the analysis and annotation of DNA microarray data. Contact: gerard.ramstein@univ-nantes.fr Supplementary information: Supplementary data are available at Bioinformatics online.

Molecular risk stratification in advanced heart failure patients

Risk stratification in advanced heart failure (HF) is crucial for the individualization of therapeutic strategy, in particular for heart transplantation and ventricular assist device implantation. We tested the hypothesis that cardiac gene expression profiling can distinguish between HF patients with different disease severity. We obtained tissue samples from both left (LV) and right (RV) ventricle of explanted hearts of 44 patients undergoing cardiac transplantation or ventricular assist device placement. Gene expression profiles were obtained using an in‐house microarray containing 4217 muscular organ‐relevant genes. Based on their clinical status, patients were classified into three HF‐severity groups: deteriorating (n= 12), intermediate (n= 19) and stable (n= 13). Two‐class statistical analysis of gene expression profiles of deteriorating and stable patients identified a 170‐gene and a 129‐gene predictor for LV and RV samples, respectively. The LV molecular predictor identified patients with stable and deteriorating status with a sensitivity of 88% and 92%, and a specificity of 100% and 96%, respectively. The RV molecular predictor identified patients with stable and deteriorating status with a sensitivity of 100% and 96%, and a specificity of 100% and 100%, respectively. The molecular prediction was reproducible across biological replicates in LV and RV samples. Gene expression profiling has the potential to reproducibly detect HF patients with highest HF severity with high sensitivity and specificity. In addition, not only LV but also RV samples could be used for molecular risk stratification with similar predictive power.

Multi-objective constrained design of nickel-base superalloys using data mining- and thermodynamics-driven genetic algorithms.

A new computational framework for systematic and optimal alloy design is introduced. It is based on a multi-objective genetic algorithm which allows (i) the screening of vast compositional ranges and (ii) the optimisation of the performance of novel alloys. Alloys performance is evaluated on the basis of their predicted constitutional and thermomechanical properties. To this end, the CALPHAD method is used for assessing equilibrium characteristics (such as constitution, stability or processability) while Gaussian processes provide an estimate of thermomechanical properties (such as tensile strength or creep resistance), based on a multi-variable non-linear regression of existing data. These three independently well-assessed tools were unified within a single C++ routine. The method was applied to the design of affordable nickel-base superalloys for service in power plants, providing numerous candidates with superior expected microstructural stability and strength. An overview of the metallurgy of optimised alloys, as well as two detailed examples of optimal alloys, suggest that improvements over current commercial alloys are achievable at lower costs.

Relapsed diffuse large B-cell lymphoma present different genomic profiles between early and late relapses

Despite major advances in first-line treatment, a significant proportion of patients with diffuse large B-cell lymphoma (DLBCL) will experience treatment failure. Prognosis is particularly poor for relapses occurring less than one year after the end of first-line treatment (early relapses/ER) compared to those occurring more than one year after (late relapses/LR). To better understand genomic alterations underlying the delay of relapse, we identified copy number variations (CNVs) on 39 tumor samples from a homogeneous series of patients included in the Collaborative Trial in Relapsed Aggressive Lymphoma (CORAL) prospective study. To identify CNVs associated with ER or LR, we devised an original method based on Significance Analysis of Microarrays, a permutation-based method which allows control of false positives due to multiple testing. Deletions of CDKN2A/B (28%) and IBTK (23%) were frequent events in relapsed DLBCLs. We identified 56 protein-coding genes and 25 long non-coding RNAs with significantly differential CNVs distribution between ER and LR DLBCLs, with a false discovery rate < 0.05. In ER DLBCLs, CNVs were related to transcription regulation, cell cycle and apoptosis, with duplications of histone H1T (31%), deletions of DIABLO (26%), PTMS (21%) and CK2B (15%). In LR DLBCLs, CNVs were related to immune response, with deletions of B2M (20%) and CD58 (10%), cell proliferation regulation, with duplications of HES1 (25%) and DVL3 (20%), and transcription regulation, with MTERF4 deletions (20%). This study provides new insights into the genetic aberrations in relapsed DLBCLs and suggest pathway-targeted therapies in ER and LR DLBCLs.

A Grammatical Swarm for protein classification

We present a grammatical swarm (GS) for the optimization of an aggregation operator. This combines the results of several classifiers into a unique score, producing an optimal ranking of the individuals. We apply our method to the identification of new members of a protein family. Support vector machine and naive Bayes classifiers exploit complementary features to compute probability estimates. A great advantage of the GS is that it produces an understandable algorithm revealing the interest of the classifiers. Due to the large volume of candidate sequences, ranking quality is of crucial importance. Consequently, our fitness criterion is based on the area under the ROC curve rather than on classification error rate. We discuss the performances obtained for a particular family, the cytokines and show that this technique is an efficient means of ranking the protein sequences.

Detection of Remote Protein Homologs Using Social Programming

We present a Grammatical Swarm (GS) for the optimization of an aggregation operator. This combines the results of several classifiers into a unique score, producing an optimal ranking of the individuals. We apply our method to the identification of new members of a protein family. Support Vector Machine and Naive Bayes classifiers exploit complementary features to compute probability estimates. A great advantage of the GS is that it produces an understandable algorithm revealing the interest of the classifiers. Due to the large volume of candidate sequences, ranking quality is of crucial importance. Consequently, our fitness criterion is based on the Area Under the ROC Curve rather than on classification error rate. We discuss the performances obtained for a particular family, the cytokines and show that this technique is an efficient means of ranking the protein sequences.

SVM-Based Classification of Distant Proteins Using Hierarchical Motifs

This article presents a discriminative approach to the protein classification in the particular case of remote homology. The protein family is modelled by a set M of motifs related to the physicochemical properties of the residues. We propose an algorithm for discovering motifs based on the ascending hierarchical classification paradigm. The set M defines a feature space of the sequences: each sequence is transformed into a vector that indicates the possible presence of the motifs belonging to M. We then use the SVM learning method to discriminate the target family. Our hierarchical motif set specifically modelises interleukins among all the structural families of the SCOP database. Our method yields a significantly better remote protein classification compared to spectrum kernel techniques.

VEGF 121 , is predictor for survival in activated B-cell-like diffuse large B-cell lymphoma and is related to an immune response gene signature conserved in cancers

Tumor microenvironment including endothelial and immune cells plays a crucial role in tumor progression and has been shown to dramatically influence cancer survival. In this study, we investigated the clinical relevance of the gene expression of key mediators of angiogenesis, VEGF isoforms 121, 165, and 189, and their receptors (VEGFR-1 and R-2) in a cohort of patients (n = 37) with relapsed/refractory diffuse large B-cell lymphoma (DLBCL) from the Collaborative Trial in Relapsed Aggressive Lymphoma (CORAL). In patients with ABC-like DLBCL, but not in patients with GCB-like DLBCL, low VEGF121 expression was associated with a significantly better survival than in those with high VEGF121 level: 4-year overall survival at 100% vs 36% (p = .011), respectively. A specific gene signature including 57 genes was correlated to VEGF121 expression level and was analyzed using a discovery process in 1,842 GSE datasets of public microarray studies. This gene signature was significantly expressed in other cancer datasets and was associated with immune response. In conclusion, low VEGF121 expression level was significantly associated with a good prognosis in relapsed/refractory ABC-like DLBCL, and with a well-conserved gene-expression profiling signature related to immune response. These findings pave the way for rationalization of drugs targeting immune response in refractory/relapsed ABC-like DLBCL.Tumor microenvironment including endothelial and immune cells plays a crucial role in tumor progression and has been shown to dramatically influence cancer survival. In this study, we investigated the clinical relevance of the gene expression of key mediators of angiogenesis, VEGF isoforms 121, 165, and 189, and their receptors (VEGFR-1 and R-2) in a cohort of patients (n = 37) with relapsed/refractory diffuse large B-cell lymphoma (DLBCL) from the Collaborative Trial in Relapsed Aggressive Lymphoma (CORAL). In patients with ABC-like DLBCL, but not in patients with GCB-like DLBCL, low VEGF121 expression was associated with a significantly better survival than in those with high VEGF121 level: 4-year overall survival at 100% vs 36% (p = .011), respectively. A specific gene signature including 57 genes was correlated to VEGF121 expression level and was analyzed using a discovery process in 1,842 GSE datasets of public microarray studies. This gene signature was significantly expressed in other cancer datasets and was associated with immune response. In conclusion, low VEGF121 expression level was significantly associated with a good prognosis in relapsed/refractory ABC-like DLBCL, and with a well-conserved gene-expression profiling signature related to immune response. These findings pave the way for rationalization of drugs targeting immune response in refractory/relapsed ABC-like DLBCL.