Analyzing the dependence structure of microarray data: a copula–based approach

The main aim of this Ph.D. dissertation is the study of clustering dependent data by means of copula functions with particular emphasis on microarray data. Copula functions are a popular multivariate modeling tool in each field where the multivariate dependence is of great interest and their use in clustering has not been still investigated. The first part of this work contains the review of the literature of clustering methods, copula functions and microarray experiments. The attention focuses on the K–means (Hartigan, 1975; Hartigan and Wong, 1979), the hierarchical (Everitt, 1974) and the model–based (Fraley and Raftery, 1998, 1999, 2000, 2007) clustering techniques because their performance is compared. Then, the probabilistic interpretation of the Sklar’s theorem (Sklar’s, 1959), the estimation methods for copulas like the Inference for Margins (Joe and Xu, 1996) and the Archimedean and Elliptical copula families are presented. In the end, applications of clustering methods and copulas to the genetic and microarray experiments are highlighted. The second part contains the original contribution proposed. A simulation study is performed in order to evaluate the performance of the K–means and the hierarchical bottom–up clustering methods in identifying clusters according to the dependence structure of the data generating process. Different simulations are performed by varying different conditions (e.g., the kind of margins (distinct, overlapping and nested) and the value of the dependence parameter ) and the results are evaluated by means of different measures of performance. In light of the simulation results and of the limits of the two investigated clustering methods, a new clustering algorithm based on copula functions (‘CoClust’ in brief) is proposed. The basic idea, the iterative procedure of the CoClust and the description of the written R functions with their output are given. The CoClust algorithm is tested on simulated data (by varying the number of clusters, the copula models, the dependence parameter value and the degree of overlap of margins) and is compared with the performance of model–based clustering by using different measures of performance, like the percentage of well–identified number of clusters and the not rejection percentage of H0 on . It is shown that the CoClust algorithm allows to overcome all observed limits of the other investigated clustering techniques and is able to identify clusters according to the dependence structure of the data independently of the degree of overlap of margins and the strength of the dependence. The CoClust uses a criterion based on the maximized log–likelihood function of the copula and can virtually account for any possible dependence relationship between observations. Many peculiar characteristics are shown for the CoClust, e.g. its capability of identifying the true number of clusters and the fact that it does not require a starting classification. Finally, the CoClust algorithm is applied to the real microarray data of Hedenfalk et al. (2001) both to the gene expressions observed in three different cancer samples and to the columns (tumor samples) of the whole data matrix.

New approaches to open problems in gene expression microarray data

In the past decade, the advent of efficient genome sequencing tools and high-throughput experimental biotechnology has lead to enormous progress in the life science. Among the most important innovations is the microarray tecnology. It allows to quantify the expression for thousands of genes simultaneously by measurin the hybridization from a tissue of interest to probes on a small glass or plastic slide. The characteristics of these data include a fair amount of random noise, a predictor dimension in the thousand, and a sample noise in the dozens. One of the most exciting areas to which microarray technology has been applied is the challenge of deciphering complex disease such as cancer. In these studies, samples are taken from two or more groups of individuals with heterogeneous phenotypes, pathologies, or clinical outcomes. these samples are hybridized to microarrays in an effort to find a small number of genes which are strongly correlated with the group of individuals. Eventhough today methods to analyse the data are welle developed and close to reach a standard organization (through the effort of preposed International project like Microarray Gene Expression Data -MGED- Society [1]) it is not unfrequant to stumble in a clinician's question that do not have a compelling statistical method that could permit to answer it.The contribution of this dissertation in deciphering disease regards the development of new approaches aiming at handle open problems posed by clinicians in handle specific experimental designs. In Chapter 1 starting from a biological necessary introduction, we revise the microarray tecnologies and all the important steps that involve an experiment from the production of the array, to the quality controls ending with preprocessing steps that will be used into the data analysis in the rest of the dissertation. While in Chapter 2 a critical review of standard analysis methods are provided stressing most of problems that In Chapter 3 is introduced a method to adress the issue of unbalanced design of miacroarray experiments. In microarray experiments, experimental design is a crucial starting-point for obtaining reasonable results. In a two-class problem, an equal or similar number of samples it should be collected between the two classes. However in some cases, e.g. rare pathologies, the approach to be taken is less evident. We propose to address this issue by applying a modified version of SAM [2]. MultiSAM consists in a reiterated application of a SAM analysis, comparing the less populated class (LPC) with 1,000 random samplings of the same size from the more populated class (MPC) A list of the differentially expressed genes is generated for each SAM application. After 1,000 reiterations, each single probe given a "score" ranging from 0 to 1,000 based on its recurrence in the 1,000 lists as differentially expressed. The performance of MultiSAM was compared to the performance of SAM and LIMMA [3] over two simulated data sets via beta and exponential distribution. The results of all three algorithms over low- noise data sets seems acceptable However, on a real unbalanced two-channel data set reagardin Chronic Lymphocitic Leukemia, LIMMA finds no significant probe, SAM finds 23 significantly changed probes but cannot separate the two classes, while MultiSAM finds 122 probes with score >300 and separates the data into two clusters by hierarchical clustering. We also report extra-assay validation in terms of differentially expressed genes Although standard algorithms perform well over low-noise simulated data sets, multi-SAM seems to be the only one able to reveal subtle differences in gene expression profiles on real unbalanced data. In Chapter 4 a method to adress similarities evaluation in a three-class prblem by means of Relevance Vector Machine [4] is described. In fact, looking at microarray data in a prognostic and diagnostic clinical framework, not only differences could have a crucial role. In some cases similarities can give useful and, sometimes even more, important information. The goal, given three classes, could be to establish, with a certain level of confidence, if the third one is similar to the first or the second one. In this work we show that Relevance Vector Machine (RVM) [2] could be a possible solutions to the limitation of standard supervised classification. In fact, RVM offers many advantages compared, for example, with his well-known precursor (Support Vector Machine - SVM [3]). Among these advantages, the estimate of posterior probability of class membership represents a key feature to address the similarity issue. This is a highly important, but often overlooked, option of any practical pattern recognition system. We focused on Tumor-Grade-three-class problem, so we have 67 samples of grade I (G1), 54 samples of grade 3 (G3) and 100 samples of grade 2 (G2). The goal is to find a model able to separate G1 from G3, then evaluate the third class G2 as test-set to obtain the probability for samples of G2 to be member of class G1 or class G3. The analysis showed that breast cancer samples of grade II have a molecular profile more similar to breast cancer samples of grade I. Looking at the literature this result have been guessed, but no measure of significance was gived before.

Identification of surface protein complexes of Streptococcus pyogenes through protein microarray technology

A systematic characterization of the composition and structure of the bacterial cell-surface proteome and its complexes can provide an invaluable tool for its comprehensive understanding. The knowledge of protein complexes composition and structure could offer new, more effective targets for a more specific and consequently effective immune response against a complex instead of a single protein. Large-scale protein-protein interaction screens are the first step towards the identification of complexes and their attribution to specific pathways. Currently, several methods exist for identifying protein interactions and protein microarrays provide the most appealing alternative to existing techniques for a high throughput screening of protein-protein interactions in vitro under reasonably straightforward conditions. In this study approximately 100 proteins of Group A Streptococcus (GAS) predicted to be secreted or surface exposed by genomic and proteomic approaches were purified in a His-tagged form and used to generate protein microarrays on nitrocellulose-coated slides. To identify protein-protein interactions each purified protein was then labeled with biotin, hybridized to the microarray and interactions were detected with Cy3-labelled streptavidin. Only reciprocal interactions, i. e. binding of the same two interactors irrespective of which of the two partners is in solid-phase or in solution, were taken as bona fide protein-protein interactions. Using this approach, we have identified 20 interactors of one of the potent toxins secreted by GAS and known as superantigens. Several of these interactors belong to the molecular chaperone or protein folding catalyst families and presumably are involved in the secretion and folding of the superantigen. In addition, a very interesting interaction was found between the superantigen and the substrate binding subunit of a well characterized ABC transporter. This finding opens a new perspective on the current understanding of how superantigens are modified by the bacterial cell in order to become major players in causing disease.

Genomic characterization of the italian wolf (Canis lupus): the genes involved in black coat colour determination and application of microarray technique for snps detection.

This study provides a comprehensive genetic overview on the endangered Italian wolf population. In particular, it focuses on two research lines. On one hand, we focalised on melanism in wolf in order to isolate a mutation related with black coat colour in canids. With several reported black individuals (an exception at European level), the Italian wolf population constituted a challenging research field posing many unanswered questions. As found in North American wolf, we reported that melanism in the Italian population is caused by a different melanocortin pathway component, the K locus, in which a beta-defensin protein acts as an alternative ligand for the Mc1r. This research project was conducted in collaboration with Prof. Gregory Barsh, Department of Genetics and Paediatrics, Stanford University. On the other hand, we performed analysis on a high number of SNPs thanks to a customized Canine microarray useful to integrate or substitute the STR markers for genotyping individuals and detecting wolf-dog hybrids. Thanks to DNA microchip technology, we obtained an impressive amount of genetic data which provides a solid base for future functional genomic studies. This study was undertaken in collaboration with Prof. Robert K. Wayne, Department of Ecology and Evolutionary Biology, University of California, Los Angeles (UCLA).

Exploring host-pathogen interactions through protein microarray. Large-scale protein microarray analysis revealed novel human receptors for the staphylococcal immune evasion protein FLIPr and for the neisserial adhesin NadA

Adhesion, immune evasion and invasion are key determinants during bacterial pathogenesis. Pathogenic bacteria possess a wide variety of surface exposed and secreted proteins which allow them to adhere to tissues, escape the immune system and spread throughout the human body. Therefore, extensive contacts between the human and the bacterial extracellular proteomes take place at the host-pathogen interface at the protein level. Recent researches emphasized the importance of a global and deeper understanding of the molecular mechanisms which underlie bacterial immune evasion and pathogenesis. Through the use of a large-scale, unbiased, protein microarray-based approach and of wide libraries of human and bacterial purified proteins, novel host-pathogen interactions were identified. This approach was first applied to Staphylococcus aureus, cause of a wide variety of diseases ranging from skin infections to endocarditis and sepsis. The screening led to the identification of several novel interactions between the human and the S. aureus extracellular proteomes. The interaction between the S. aureus immune evasion protein FLIPr (formyl-peptide receptor like-1 inhibitory protein) and the human complement component C1q, key players of the offense-defense fighting, was characterized using label-free techniques and functional assays. The same approach was also applied to Neisseria meningitidis, major cause of bacterial meningitis and fulminant sepsis worldwide. The screening led to the identification of several potential human receptors for the neisserial adhesin A (NadA), an important adhesion protein and key determinant of meningococcal interactions with the human host at various stages. The interaction between NadA and human LOX-1 (low-density oxidized lipoprotein receptor) was confirmed using label-free technologies and cell binding experiments in vitro. Taken together, these two examples provided concrete insights into S. aureus and N. meningitidis pathogenesis, and identified protein microarray coupled with appropriate validation methodologies as a powerful large scale tool for host-pathogen interactions studies.

ABC transporters of the A subfamily: potential prognostic markers and candidates for antibody selection from phage-display libraries for therapeutic use in Ewing sarcoma

The prognostic value of ABC transporters in Ewing sarcoma is still poorly explored and controversial. We described for the first time the impact of various ABCs on Ewing sarcoma prognosis by assessment of their gene expression in two independent cohorts of patients. Unexpected associations with favourable outcomes were observed for two ABCs of the A-subfamily, ABCA6 and ABCA7, whereas no associations with the canonical multidrug ABC transporters were identified. The ABCs of the A-subfamily are involved in cholesterol/phospholipids transportation and efflux from cells. Our clinical data support the drug-efflux independent contribution to cancer progression of the ABCAs, which has been confirmed in PDX-derived cell lines. The impact of these ABCA transporters on tumor progression seems to be mediated by lowering intracellular cholesterol, supporting the role of these proteins in lipid transport. In addition, the gene expression of ABCA6 and ABCA7 is regulated by transcription factors which control lipid metabolism: ABCA6 was induced by the binding of FoxO1/FoxO3a to its promoter and repressed by IGF1R/Akt signaling, whereas the expression of ABCA7 was regulated by p53. The data point to ABCA6 and ABCA7 as potential prognostic markers in Ewing sarcoma and suggest the IGF1/ABCA/lipid axis as an intriguing therapeutic target. Agonist monoclonal antibodies towards ABCA6/7 or inhibitors of cholesterol biosynthesis, such as statins or aminobiphoshonates, may be investigated as therapeutic options in combination with chemotherapy. Considering that no monoclonal antibodies selectively targeting extracellular domains of ABCA6/7 are available, the second part of the project has been dedicated to the generation of human antibody phage-display libraries as tools for selecting monoclonal antibodies. A novel synthetic human antibody phage-display library has been designed, cloned and characterized. The library takes advantages of the high variability of a designed naïve repertoire to be a useful tool for isolating antibodies towards all potential antigens, including the ABCAs.

Interrogation of human monoclonal antibodies induced by 4C-MenB to identify protective antigens contained in the OMV component

Neisseria meningitidis is a gram negative human obligated pathogen, mostly found as a commensal in the oropharyngeal mucosa of healthy individuals. It can invade this epithelium determining rare but devastating and fast progressing outcomes, such as meningococcal meningitidis and septicemia, leading to death (about 135000 per year worldwide). Conjugated vaccines for serogroups A, C, W135, X and Y were developed, while for N. meningitidis serogroup B (MenB) the vaccines were based on Outern Membrane Vesicles (OMV). One of them is the 4C-MenB (Bexsero). The antigens included in this vaccine’s formulation are, in addition to the OMV from New Zeland epidemic strain 98/254, three recombinant proteins: NadA, NHBA and fHbp. While the role of these recombinant components was deeply characterized, the vesicular contribution in 4C-MenB elicited protection is mediated mainly by porin A and other unidentified antigens. To unravel the relative contribution of these different antigens in eliciting protective antibody responses, we isolated human monoclonal antibodies (mAbs) from single-cell sorted plasmablasts of 3 adult vaccinees peripheral blood. mAbs have been screened for binding to 4C-MenB components by Luminex bead-based assay. OMV-specific mAbs were purified and tested for functionality by serum bactericidal assay (SBA) on 18 different MenB strains and characterized in a protein microarray containing a panel of prioritized meningococcal proteins. The bactericidal mAbs identified to recognize the outer membrane proteins PorA and PorB, stating the importance of PorB in cross-strain protection. In addition, RmpM, BamE, Hyp1065 and ComL were found as immunogenic components of the 4C-MenB vaccine.

Neonatal outcomes following maternal SARS-CoV-2 infection in pregnancy - longitudinal follow-up and assessment of transplacental antibody transfer

BACKGROUND: Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection in pregnancy has been associated with multiple adverse pregnancy outcomes, including the risk of in utero mother-to-child transmission. Short- and long-term outcomes of SARS-CoV-2 exposed neonates and the extent to which maternal SARS-CoV-2 antibodies are transferred to neonates are still unclear. METHODS: Prospective observational study enrolling neonates born to mothers with SARS-CoV-2 infection in pregnancy, between April 2020-April 2021. Neonates were evaluated at birth and enrolled in a 12-month follow-up. SARS-CoV-2 IgG transplacental transfer ratio was assessed in mother-neonate dyads at birth. Maternal derived IgG were followed in infants until negativizing. RESULTS: Of 2745 neonates, 106 (3.9%) were delivered by mothers with SARS-CoV-2 infection in pregnancy. Seventy-six of 106 (71.7%) mothers were symptomatic. Median gestational age and mean birth weight were 39 weeks (range 25+5-41+4) and 3305 grams (SD 468). Six of 106 (6%) neonates were born preterm, without significant differences between asymptomatic and symptomatic mothers (P=0.67). No confirmed cases of in utero infection were detected. All infants had normal cerebral ultrasound and clinical evaluation at birth and during follow-up, until a median age of 7 months (range 5-12). All mothers and 96/106 (90.5%) neonates had detectable SARS-CoV-2 IgG at birth. Transplacental transfer ratio was higher following second trimester maternal infections (mean 0.940.46 versus 1.070.64 versus 0.750.44, P=0.039), but was not significantly different between asymptomatic and symptomatic women (P=0.20). IgG level in infants progressively decreased after birth: at 3 months 53% (51/96) and at four months 68% (63/96) had lost maternal antibodies respectively. The durability of maternal antibodies was positively correlated to the IgG level at birth (r=0.66; P<0.00001). CONCLUSIONS: Maternal SARS-CoV-2 infection was not associated with increased neonatal or long-term morbidity. No cases of confirmed in utero infection were detected. Efficient transplacental IgG transfer was found following second trimester maternal infections.

Research of predictive biomarkers to anti-CD22 antibody-drug conjugate treatment in B-cell acute lymphoblastic leukemia

Background: The treatment of B-cell acute lymphoblastic leukemia (B-ALL) has been enriched by novel agents targeting surface markers CD19 and CD22. Inotuzumab ozogamicin (INO) is a CD22-calicheamicin conjugated monoclonal antibody approved in the setting of relapse/refractory (R/R) B-ALL able to induce a high rate of deep responses, not durable over time. Aims: This study aims to identify predictive biomarkers to INO treatment in B- ALL by flow cytometric analysis of CD22 expression and gene expression profile. Materials and methods: Firstly, the impact on patient outcome in 30 R/R B-ALL patients of baseline CD22 expression in terms of CD22 blast percentage and CD22 fluorescent intensity (CD22-FI) was explored. Secondly, baseline gene expression profile of 18 R/R B-ALL patient samples was analyzed. For statistical analysis of differentially expressed genes (DEGs) patients were divided in non-responders (NR), defined as either INO-refractory or with duration of response (DoR) < 3 months, and responders (R). Gene expression results were analyzed with Ingenuity pathway analysis (IPA). Results: In our patient set higher CD22-FI, defined as higher quartiles (Q2-Q4), correlated with better patient outcome in terms of CR rate, OS and DoR, compared to lower CD22-FI (Q1). CD22 blast percentage was less able to discriminate patients’ outcome, although a trend for better outcome in patients with CD22 ≥ 90% could be appreciated. Concerning gene expression profile, 32 genes with corrected p value <0.05 and absolute FC ≥2 were differentially expressed in NR as compared to R. IPA upstream regulator and regulator effect analysis individuated the inhibition of tumor suppressor HIPK2 as causal upstream condition of the downregulation of 6 DEGs. Conclusions: CD22-FI integrates CD22-percentage on leukemic blasts for a more comprehensive target pre-treatment evaluation. Moreover, a unique pattern of gene expression signature based on HIPK2 downregulation was identified, providing important insights in mechanisms of resistance to INO.