Advanced Analysis and Validation of a microRNA Signature for Fanconi Anemia.

Some years ago, we reported the generation of a Fanconi anemia (FA) microRNA signature. This study aims to develop an analytical strategy to select a smaller and more reliable set of molecules that could be tested for potential benefits for the FA phenotype, elucidate its biochemical and molecular mechanisms, address experimental activity, and evaluate its possible impact on FA therapy. In silico analyses of the data obtained in the original study were thoroughly processed and anenrichment analysis was employed to identify the classes of genes that are over-represented in the FA-miRNA population under study. Primary bone marrow mononuclear cells (MNCs) from sixFA patients and sixhealthy donors as control samples were employed in the study. RNAs containing the small RNA fractions were reverse-transcribed and real-time PCR was performed in triplicate using the specific primers. Experiments were performed in triplicate.The in-silico analysis reported six miRNAs as likely contributors to the complex pathological spectrum of FA. Among these, three miRNAs were validated by real-time PCR. Primary bone marrow mononuclear cells (MNCs) reported a significant reduction in the expression level of miRNA-1246 and miRNA-206 in the FA samples in comparison to controls.This study highlights several biochemical pathways as culprits in the phenotypic manifestations and the pathophysiological mechanisms acting in FA. A relatively low number of miRNAs appear involved in all these different phenotypes, demonstrating the extreme plasticity of the gene expression modulation. This study further highlights miR-206 as a pivotal player in regulatory functions and signaling in the bone marrow mesenchymal stem cell (BMSC) process in FA. Due to this evidence, the activity of miR-206 in FA deserves specific experimental scrutiny. The results, here presented, might be relevant in the management of FA.

Gut-microbiota in children and adolescents with obesity: inferred functional analysis and machine-learning algorithms to classify microorganisms.

The fecal microbiome of 55 obese children and adolescents (BMI-SDS 3.2 ± 0.7) and of 25 normal-weight subjects, matched both for age and sex (BMI-SDS - 0.3 ± 1.1) was analysed. Streptococcus, Acidaminococcus, Sutterella, Prevotella, Sutterella wadsworthensis, Streptococcus thermophilus, and Prevotella copri positively correlated with obesity. The inferred pathways strongly associated with obesity concern the biosynthesis pathways of tyrosine, phenylalanine, tryptophan and methionine pathways. Furthermore, polyamine biosynthesis virulence factors and pro-inflammatory lipopolysaccharide biosynthesis pathway showed higher abundances in obese samples, while the butanediol biosynthesis showed low abundance in obese subjects. Different taxa strongly linked with obesity have been related to an increased risk of multiple diseases involving metabolic pathways related to inflammation (polyamine and lipopolysaccharide biosynthesis). Cholesterol, LDL, and CRP positively correlated with specific clusters of microbial in obese patients. The Firmicutes/Bacteroidetes-ratio was lower in obese samples than in controls and differently from the literature we state that this ratio could not be a biomarker for obesity.

An overview of data integration in neuroscience with focus on Alzheimer's Disease.

This work represents the first attempt to provide an overview of how to face data integration as the result of a dialogue between neuroscientists and computer scientists. Indeed, data integration is fundamental for studying complex multifactorial diseases, such as the neurodegenerative diseases. This work aims at warning the readers of common pitfalls and critical issues in both medical and data science fields. In this context, we define a road map for data scientists when they first approach the issue of data integration in the biomedical domain, highlighting the challenges that inevitably emerge when dealing with heterogeneous, large-scale and noisy data and proposing possible solutions. Here, we discuss data collection and statistical analysis usually seen as parallel and independent processes, as cross-disciplinary activities. Finally, we provide an exemplary application of data integration to address Alzheimer's Disease (AD), which is the most common multifactorial form of dementia worldwide. We critically discuss the largest and most widely used datasets in AD, and demonstrate how the emergence of machine learning and deep learning methods has had a significant impact on disease's knowledge particularly in the perspective of an early AD diagnosis.

Effects of Chronic Arginase Inhibition with Norvaline on Tau Pathology and Brain Glucose Metabolism in Alzheimer's Disease Mice.

Alzheimer's disease (AD) is an insidious neurodegenerative disorder representing a serious continuously escalating medico-social problem. The AD-associated progressive dementia is followed by gradual formation of amyloid plaques and neurofibrillary tangles in the brain. Though, converging evidence indicates apparent metabolic dysfunctions as key AD characteristic. In particular, late-onset AD possesses a clear metabolic signature. Considerable brain insulin signaling impairment and a decline in glucose metabolism are common AD attributes. Thus, positron emission tomography (PET) with glucose tracers is a reliable non-invasive tool for early AD diagnosis and treatment efficacy monitoring. Various approaches and agents have been trialed to modulate insulin signaling. Accumulating data point to arginase inhibition as a promising direction to treat AD via diverse molecular mechanisms involving, inter alia, the insulin pathway. Here, we use a transgenic AD mouse model, demonstrating age-dependent brain insulin signaling abnormalities, reduced brain insulin receptor levels, and substantial energy metabolism alterations, to evaluate the effects of arginase inhibition with Norvaline on glucose metabolism. We utilize fluorodeoxyglucose whole-body micro-PET to reveal a significant treatment-associated increase in glucose uptake by the brain tissue in-vivo. Additionally, we apply advanced molecular biology and bioinformatics methods to explore the mechanisms underlying the effects of Norvaline on glucose metabolism. We demonstrate that treatment-associated improvement in glucose utilization is followed by significantly elevated levels of insulin receptor and glucose transporter-3 expression in the mice hippocampi. Additionally, Norvaline diminishes the rate of Tau protein phosphorylation. Our results suggest that Norvaline interferes with AD pathogenesis. These findings open new avenues for clinical evaluation and innovative drug development.

Pathways and microbiome modifications related to surgery and enterocolitis in Hirschsprung disease.

BACKGROUND: Hirschsprung disease (HSCR) is a congenital anomaly of the enteric nervous system. Abnormal microbiome composition was reported in HSCR patients. In this study, we addressed and analyzed microbiome modifications with relation tosurgery and HSCR associated enterocolitis (HAEC). METHODS: The faecal microbiome of 31 HSCR patients (overall 64 samples) was analyzed. HAEC was diagnosed and classified according to a combination of Pastor's and Elhalabi's criteria. Stool samples were analyzed by 16S sequencing (7 out of 9 polymorphic regions). Compositional and relative abundance profiles, as well as the functional potentials of the microbial community, were analyzed with the marker gene sequencing profiles using PICRUSt. RESULTS: The relative abundance of Bacteroidetes showed a severe decrease with slow recovery after surgery. Conversely, Proteobacteria transiently increased their abundance. Noteworthy, a strong linkage has been found between Proteobacteria descendants and HAEC occurrences. The inferred functional analysis indicated that virulence factors and fimbriae or pili might be associated with HAEC. CONCLUSIONS: Our study, addressing microbiome dynamics, demonstrated relevant changes after surgical manipulation. Alpha-diversity analyses indicated that surgery deeply affects microbiome composition. Proteobacteria and Enterobacteriaceae seem to play a pivotal role in HAEC occurrences. Several virulence factors, such as fimbriae or pili, might explain the HAEC-predisposing potential of selected microbiomes. These results suggest some innovative therapeutic approaches that deserve to be tested in appropriate clinical trials.

Case Report: Whole Exome Sequencing Revealed Disease-Causing Variants in Two Genes in a Patient With Autism Spectrum Disorder, Intellectual Disability, Hyperactivity, Sleep and Gastrointestinal Disturbances.

Autism Spectrum Disorder (ASD) refers to a broad range of conditions characterized by difficulties in communication, social interaction and behavior, and may be accompanied by other medical or psychiatric conditions. Patients with ASD and comorbidities are often difficult to diagnose because of the tendency to consider the multiple symptoms as the presentation of a complicated syndromic form. This view influences variant filtering which might ignore causative variants for specific clinical features shown by the patient. Here we report on a male child diagnosed with ASD, showing cognitive and motor impairments, stereotypies, hyperactivity, sleep, and gastrointestinal disturbances. The analysis of whole exome sequencing (WES) data with bioinformatic tools for oligogenic diseases helped us to identify two major previously unreported pathogenetic variants: a maternally inherited missense variant (p.R4122H) in HUWE1, an ubiquitin protein ligase associated to X-linked intellectual disability and ASD; and a de novo stop variant (p.Q259X) in TPH2, encoding the tryptophan hydroxylase 2 enzyme involved in serotonin synthesis and associated with susceptibility to attention deficit-hyperactivity disorder (ADHD). TPH2, expressed in central and peripheral nervous tissues, modulates various physiological functions, including gut motility and sleep. To the best of our knowledge, this is the first case presenting with ASD, cognitive impairment, sleep, and gastrointestinal disturbances linked to both HUWE1 and TPH2 genes. Our findings could contribute to the existing knowledge on clinical and genetic diagnosis of patients with ASD presentation with comorbidities.

A telescope GWAS analysis strategy, based on SNPs-genes-pathways ensamble and on multivariate algorithms, to characterize late onset Alzheimer's disease.

Genome-wide association studies (GWAS) have revealed a plethora of putative susceptibility genes for Alzheimer's disease (AD), with the sole exception of APOE gene unequivocally validated in independent study. Considering that the etiology of complex diseases like AD could depend on functional multiple genes interaction network, here we proposed an alternative GWAS analysis strategy based on (i) multivariate methods and on a (ii) telescope approach, in order to guarantee the identification of correlated variables, and reveal their connections at three biological connected levels. Specifically as multivariate methods, we employed two machine learning algorithms and a genetic association test and we considered SNPs, Genes and Pathways features in the analysis of two public GWAS dataset (ADNI-1 and ADNI-2). For each dataset and for each feature we addressed two binary classifications tasks: cases vs. controls and the low vs. high risk of developing AD considering the allelic status of APOEe4. This complex strategy allowed the identification of SNPs, genes and pathways lists statistically robust and meaningful from the biological viewpoint. Among the results, we confirm the involvement of TOMM40 gene in AD and we propose GRM7 as a novel gene significantly associated with AD.

Gut Microbiota in T1DM-Onset Pediatric Patients: Machine-Learning Algorithms to Classify Microorganisms as Disease Linked.

AIMS: The purpose of this work is to find the gut microbial fingerprinting of pediatric patients with type 1 diabetes. METHODS: The microbiome of 31 children with type 1 diabetes at onset and of 25 healthy children was determined using multiple polymorphic regions of the 16S ribosomal RNA. We performed machine-learning analyses and metagenome functional analysis to identify significant taxa and their metabolic pathways content. RESULTS: Compared with healthy controls, patients showed a significantly higher relative abundance of the following most important taxa: Bacteroides stercoris, Bacteroides fragilis, Bacteroides intestinalis, Bifidobacterium bifidum, Gammaproteobacteria and its descendants, Holdemania, and Synergistetes and its descendants. On the contrary, the relative abundance of Bacteroides vulgatus, Deltaproteobacteria and its descendants, Parasutterella and the Lactobacillus, Turicibacter genera were significantly lower in patients with respect to healthy controls. The predicted metabolic pathway more associated with type 1 diabetes patients concerns "carbon metabolism," sugar and iron metabolisms in particular. Among the clinical variables considered, standardized body mass index, anti-insulin autoantibodies, glycemia, hemoglobin A1c, Tanner stage, and age at onset emerged as most significant positively or negatively correlated with specific clusters of taxa. CONCLUSIONS: The relative abundance and supervised analyses confirmed the importance of B stercoris in type 1 diabetes patients at onset and showed a relevant role of Synergistetes and its descendants in patients with respect to healthy controls. In general the robustness and coherence of the showed results underline the relevance of studying the microbioma using multiple polymorphic regions, different types of analysis, and different approaches within each analysis.

Transplantation Induces Profound Changes in the Transcriptional Asset of Hematopoietic Stem Cells: Identification of Specific Signatures Using Machine Learning Techniques.

During the phase of proliferation needed for hematopoietic reconstitution following transplantation, hematopoietic stem/progenitor cells (HSPC) must express genes involved in stem cell self-renewal. We investigated the expression of genes relevant for self-renewal and expansion of HSPC (operationally defined as CD34+ cells) in steady state and after transplantation. Specifically, we evaluated the expression of ninety-one genes that were analyzed by real-time PCR in CD34+ cells isolated from (i) 12 samples from umbilical cord blood (UCB); (ii) 15 samples from bone marrow healthy donors; (iii) 13 samples from bone marrow after umbilical cord blood transplant (UCBT); and (iv) 29 samples from patients after transplantation with adult hematopoietic cells. The results show that transplanted CD34+ cells from adult cells acquire an asset very different from transplanted CD34+ cells from cord blood. Multivariate machine learning analysis (MMLA) showed that four specific gene signatures can be obtained by comparing the four types of CD34+ cells. In several, but not all cases, transplanted HSPC from UCB overexpress reprogramming genes. However, these remarkable changes do not alter the commitment to hematopoietic lineage. Overall, these results reveal undisclosed aspects of transplantation biology.

Tracing CLL-biased stereotyped immunoglobulin gene rearrangements in normal B cell subsets using a high-throughput immunogenetic approach.

BACKGROUND: B cell receptor Immunoglobulin (BcR IG) repertoire of Chronic Lymphocytic Leukemia (CLL) is characterized by the expression of quasi-identical BcR IG. These are observed in approximately 30% of patients, defined as stereotyped receptors and subdivided into subsets based on specific VH CDR3 aa motifs and phylogenetically related IGHV genes. Although relevant to CLL ontogeny, the distribution of CLL-biased stereotyped immunoglobulin rearrangements (CBS-IG) in normal B cells has not been so far specifically addressed using modern sequencing technologies. Here, we have investigated the presence of CBS-IG in splenic B cell subpopulations (s-BCS) and in CD5+ and CD5- B cells from the spleen and peripheral blood (PB). METHODS: Fractionation of splenic B cells into 9 different B cell subsets and that of spleen and PB into CD5+ and CD5- cells were carried out by FACS sorting. cDNA sequences of BcR IG gene rearrangements were obtained by NGS. Identification of amino acidic motifs typical of CLL stereotyped subsets was carried out on IGHV1-carrying gene sequences and statistical evaluation has been subsequently performed to assess stereotypes distribution. RESULTS: CBS-IG represented the 0.26% average of IGHV1 genes expressing sequences, were detected in all of the BCS investigated. CBS-IG were more abundant in splenic and circulating CD5+ B (0.57%) cells compared to CD5- B cells (0.17%). In all instances, most CBS IG did not exhibit somatic hypermutation similar to CLL stereotyped receptors. However, compared to CLL, they exhibited a different CLL subset distribution and a broader utilization of the genes of the IGHV1 family. CONCLUSIONS: CBS-IG receptors appear to represent a part of the "public" BcR repertoire in normal B cells. This repertoire is observed in all BCS excluding the hypothesis that CLL stereotyped BcR accumulate in a specific B cell subset, potentially capable of originating a leukemic clone. The different relative representation of CBS-IG in normal B cell subgroups suggests the requirement for additional selective processes before a full transformation into CLL is achieved.

Different Features of Tumor-Associated NK Cells in Patients With Low-Grade or High-Grade Peritoneal Carcinomatosis.

Peritoneal carcinomatosis (PC) is a rare disease defined as diffused implantation of neoplastic cells in the peritoneal cavity. This clinical picture occurs during the evolution of peritoneal tumors, and it is the main cause of morbidity and mortality of patients affected by these pathologies, though cytoreductive surgery with heated intra-peritoneal chemotherapy (CRS/HIPEC) is yielding promising results. In the present study, we evaluated whether the tumor microenvironment of low-grade and high-grade PC could affect the phenotypic and functional features and thus the anti-tumor potential of NK cells. We show that while in the peritoneal fluid (PF) of low-grade PC most CD56dim NK cells show a relatively immature phenotype (NKG2A+KIR-CD57-CD16dim), in the PF of high-grade PC NK cells are, in large majority, mature (CD56dimKIR+CD57+CD16bright). Furthermore, in low-grade PC, PF-NK cells are characterized by a sharp down-regulation of some activating receptors, primarily NKp30 and DNAM-1, while, in high-grade PC, PF-NK cells display a higher expression of the PD-1 inhibitory checkpoint. The compromised phenotype observed in low-grade PC patients corresponds to a functional impairment. On the other hand, in the high-grade PC patients PF-NK cells show much more important defects that only partially reflect the compromised phenotype detected. These data suggest that the PC microenvironment may contribute to tumor escape from immune surveillance by inducing different NK cell impaired features leading to altered anti-tumor activity. Notably, after CRS/HIPEC treatment, the altered NK cell phenotype of a patient with a low-grade disease and favorable prognosis was reverted to a normal one. Our present data offer a clue for the development of new immunotherapeutic strategies capable of restoring the NK-mediated anti-tumor responses in association with the CRS/HIPEC treatment to increase the effectiveness of the current therapy.

New miRNA Signature Heralds Human NK Cell Subsets at Different Maturation Steps: Involvement of miR-146a-5p in the Regulation of KIR Expression.

Natural killer cells are cytotoxic innate lymphoid cells that play an important role for early host defenses against infectious pathogens and surveillance against tumor. In humans, NK cells may be divided in various subsets on the basis of the relative CD56 expression and of the low-affinity FcγRIIIA CD16. In particular, the two main NK cell subsets are represented by the CD56bright/CD16-/dim and the CD56dim/CD16bright NK cells. Experimental evidences indicate that CD56bright and CD56dim NK cells represent different maturative stages of the NK cell developmental pathway. We identified multiple miRNAs differentially expressed in CD56bright/CD16- and CD56dim/CD16bright NK cells using both univariate and multivariate analyses. Among these, we found a few miRNAs with a consistent differential expression in the two NK cell subsets, and with an intermediate expression in the CD56bright/CD16dim NK cell subset, representing a transitional step of maturation of NK cells. These analyses allowed us to establish the existence of a miRNA signature able to efficiently discriminate the two main NK cell subsets regardless of their surface phenotype. In addition, by analyzing the putative targets of representative miRNAs we show that hsa-miR-146a-5p, may be involved in the regulation of killer Ig-like receptor (KIR) expression. These results contribute to a better understanding of the physiologic significance of miRNAs in the regulation of the development/function of human NK cells. Moreover, our results suggest that hsa-miR-146a-5p targeting, resulting in KIR down-regulation, may be exploited to generate/increment the effect of NK KIR-mismatching against HLA-class I+ tumor cells and thus improve the NK-mediated anti-tumor activity.

Enhancing Interpretability of Gene Signatures with Prior Biological Knowledge.

Biological interpretability is a key requirement for the output of microarray data analysis pipelines. The most used pipeline first identifies a gene signature from the acquired measurements and then uses gene enrichment analysis as a tool for functionally characterizing the obtained results. Recently Knowledge Driven Variable Selection (KDVS), an alternative approach which performs both steps at the same time, has been proposed. In this paper, we assess the effectiveness of KDVS against standard approaches on a Parkinson's Disease (PD) dataset. The presented quantitative analysis is made possible by the construction of a reference list of genes and gene groups associated to PD. Our work shows that KDVS is much more effective than the standard approach in enhancing the interpretability of the obtained results.

A Reassessment of IgM Memory Subsets in Humans.

From paired blood and spleen samples from three adult donors, we performed high-throughput VH sequencing of human B cell subsets defined by IgD and CD27 expression: IgD(+)CD27(+) ("marginal zone [MZ]"), IgD(-)CD27(+) ("memory," including IgM ["IgM-only"], IgG and IgA) and IgD(-)CD27(-) cells ("double-negative," including IgM, IgG, and IgA). A total of 91,294 unique sequences clustered in 42,670 clones, revealing major clonal expansions in each of these subsets. Among these clones, we further analyzed those shared sequences from different subsets or tissues for VH gene mutation, H-CDR3-length, and VH/JH usage, comparing these different characteristics with all sequences from their subset of origin for which these parameters constitute a distinct signature. The IgM-only repertoire profile differed notably from that of MZ B cells by a higher mutation frequency and lower VH4 and higher JH6 gene usage. Strikingly, IgM sequences from clones shared between the MZ and the memory IgG/IgA compartments showed a mutation and repertoire profile of IgM-only and not of MZ B cells. Similarly, all IgM clonal relationships (among MZ, IgM-only, and double-negative compartments) involved sequences with the characteristics of IgM-only B cells. Finally, clonal relationships between tissues suggested distinct recirculation characteristics between MZ and switched B cells. The "IgM-only" subset (including cells with its repertoire signature but higher IgD or lower CD27 expression levels) thus appear as the only subset showing precursor-product relationships with CD27(+) switched memory B cells, indicating that they represent germinal center-derived IgM memory B cells and that IgM memory and MZ B cells constitute two distinct entities.

Regenerative medicine for the treatment of Teno-desmic injuries of the equine. A series of 150 horses treated with platelet-derived growth factors.

BACKGROUND/AIM: The aim of the present study was to evaluate the safety and the clinical outcome of platelet-rich plasma for the treatment of teno-desmic injures in competition horses. PATIENTS AND METHODS: From January 2009 to December 2011, 150 sport horses suffering from teno-desmic injuries were treated with no-gelled platelet-concentrate. RESULTS: No horse showed any major adverse reaction as a result of the procedure. Full healing was obtained for 81% of the horses. Twelve percent had clinical improvement and only 7% a failure. Eight percent of cases of relapse were observed. No statistically significant correlation existed between clinical outcome and the area of the lesion. A statistically significant correlation existed between the clinical outcome and the age of the horse. CONCLUSION: Treatment with platelet-derived growth factors leads to the formation of a tendon with normal morphology and functionality, which translate in the resumption of the agonistic activity for the horses we treated.

Knowledge Driven Variable Selection (KDVS) - a new approach to enrichment analysis of gene signatures obtained from high-throughput data.

BACKGROUND: High-throughput (HT) technologies provide huge amount of gene expression data that can be used to identify biomarkers useful in the clinical practice. The most frequently used approaches first select a set of genes (i.e. gene signature) able to characterize differences between two or more phenotypical conditions, and then provide a functional assessment of the selected genes with an a posteriori enrichment analysis, based on biological knowledge. However, this approach comes with some drawbacks. First, gene selection procedure often requires tunable parameters that affect the outcome, typically producing many false hits. Second, a posteriori enrichment analysis is based on mapping between biological concepts and gene expression measurements, which is hard to compute because of constant changes in biological knowledge and genome analysis. Third, such mapping is typically used in the assessment of the coverage of gene signature by biological concepts, that is either score-based or requires tunable parameters as well, limiting its power. RESULTS: We present Knowledge Driven Variable Selection (KDVS), a framework that uses a priori biological knowledge in HT data analysis. The expression data matrix is transformed, according to prior knowledge, into smaller matrices, easier to analyze and to interpret from both computational and biological viewpoints. Therefore KDVS, unlike most approaches, does not exclude a priori any function or process potentially relevant for the biological question under investigation. Differently from the standard approach where gene selection and functional assessment are applied independently, KDVS embeds these two steps into a unified statistical framework, decreasing the variability derived from the threshold-dependent selection, the mapping to the biological concepts, and the signature coverage. We present three case studies to assess the usefulness of the method. CONCLUSIONS: We showed that KDVS not only enables the selection of known biological functionalities with accuracy, but also identification of new ones. An efficient implementation of KDVS was devised to obtain results in a fast and robust way. Computing time is drastically reduced by the effective use of distributed resources. Finally, integrated visualization techniques immediately increase the interpretability of results. Overall, KDVS approach can be considered as a viable alternative to enrichment-based approaches.

Effect of size and heterogeneity of samples on biomarker discovery: synthetic and real data assessment.

MOTIVATION: The identification of robust lists of molecular biomarkers related to a disease is a fundamental step for early diagnosis and treatment. However, methodologies for the discovery of biomarkers using microarray data often provide results with limited overlap. These differences are imputable to 1) dataset size (few subjects with respect to the number of features); 2) heterogeneity of the disease; 3) heterogeneity of experimental protocols and computational pipelines employed in the analysis. In this paper, we focus on the first two issues and assess, both on simulated (through an in silico regulation network model) and real clinical datasets, the consistency of candidate biomarkers provided by a number of different methods. METHODS: We extensively simulated the effect of heterogeneity characteristic of complex diseases on different sets of microarray data. Heterogeneity was reproduced by simulating both intrinsic variability of the population and the alteration of regulatory mechanisms. Population variability was simulated by modeling evolution of a pool of subjects; then, a subset of them underwent alterations in regulatory mechanisms so as to mimic the disease state. RESULTS: The simulated data allowed us to outline advantages and drawbacks of different methods across multiple studies and varying number of samples and to evaluate precision of feature selection on a benchmark with known biomarkers. Although comparable classification accuracy was reached by different methods, the use of external cross-validation loops is helpful in finding features with a higher degree of precision and stability. Application to real data confirmed these results.

A computational procedure for functional characterization of potential marker genes from molecular data: Alzheimer's as a case study.

BACKGROUND: A molecular characterization of Alzheimer's Disease (AD) is the key to the identification of altered gene sets that lead to AD progression. We rely on the assumption that candidate marker genes for a given disease belong to specific pathogenic pathways, and we aim at unveiling those pathways stable across tissues, treatments and measurement systems. In this context, we analyzed three heterogeneous datasets, two microarray gene expression sets and one protein abundance set, applying a recently proposed feature selection method based on regularization. RESULTS: For each dataset we identified a signature that was successively evaluated both from the computational and functional characterization viewpoints, estimating the classification error and retrieving the most relevant biological knowledge from different repositories. Each signature includes genes already known to be related to AD and genes that are likely to be involved in the pathogenesis or in the disease progression. The integrated analysis revealed a meaningful overlap at the functional level. CONCLUSIONS: The identification of three gene signatures showing a relevant overlap of pathways and ontologies, increases the likelihood of finding potential marker genes for AD.