VLPs generated by the fusion of RSV-F or hMPV-F glycoprotein to HIV-Gag show improved immunogenicity and neutralizing response in mice

Respiratory syncytial virus (RSV) and human metapneumovirus (hMPV) vaccines have been long overdue. Structure-based vaccine design created a new momentum in the last decade, and the first RSV vaccines have finally been approved in older adults and pregnant individuals. These vaccines are based on recombinant stabilized pre-fusion F glycoproteins administered as soluble proteins. Multimeric antigenic display could markedly improve immunogenicity and should be evaluated in the next generations of vaccines. Here we tested a new virus like particles-based vaccine platform which utilizes the direct fusion of an immunogen of interest to the structural human immunodeficient virus (HIV) protein Gag to increase its surface density and immunogenicity. We compared, in mice, the immunogenicity of RSV-F or hMPV-F based immunogens delivered either as soluble proteins or displayed on the surface of our VLPs. VLP associated F-proteins showed better immunogenicity and induced superior neutralizing responses. Moreover, when combining both VLP associated and soluble immunogens in a heterologous regimen, VLP-associated immunogens provided added benefits when administered as the prime immunization.

Immunization with V987H-stabilized Spike glycoprotein protects K18-hACE2 mice and golden Syrian hamsters upon SARS-CoV-2 infection

Safe and effective severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccines are crucial to fight against the coronavirus disease 2019 pandemic. Most vaccines are based on a mutated version of the Spike glycoprotein [K986P/V987P (S-2P)] with improved stability, yield and immunogenicity. However, S-2P is still produced at low levels. Here, we describe the V987H mutation that increases by two-fold the production of the recombinant Spike and the exposure of the receptor binding domain (RBD). S-V987H immunogenicity is similar to S-2P in mice and golden Syrian hamsters (GSH), and superior to a monomeric RBD. S-V987H immunization confer full protection against severe disease in K18-hACE2 mice and GSH upon SARS-CoV-2 challenge (D614G or B.1.351 variants). Furthermore, S-V987H immunized K18-hACE2 mice show a faster tissue viral clearance than RBD- or S-2P-vaccinated animals challenged with D614G, B.1.351 or Omicron BQ1.1 variants. Thus, S-V987H protein might be considered for future SARS-CoV-2 vaccines development.

A monoclonal antibody targeting a large surface of the receptor binding motif shows pan-neutralizing SARS-CoV-2 activity

Here we report the characterization of 17T2, a SARS-CoV-2 pan-neutralizing human monoclonal antibody isolated from a COVID-19 convalescent individual infected during the first pandemic wave. 17T2 is a class 1 VH1-58/κ3-20 antibody, derived from a receptor binding domain (RBD)-specific IgA+ memory B cell, with a broad neutralizing activity against former and new SARS-CoV-2 variants, including XBB.1.16 and BA.2.86 Omicron subvariants. Consistently, 17T2 demonstrates in vivo prophylactic and therapeutic activity against Omicron BA.1.1 infection in K18-hACE2 mice. Cryo-electron microscopy reconstruction shows that 17T2 binds the BA.1 spike with the RBD in "up" position and blocks the receptor binding motif, as other structurally similar antibodies do, including S2E12. Yet, unlike S2E12, 17T2 retains its neutralizing activity against all variants tested, probably due to a larger RBD contact area. These results highlight the impact of small structural antibody changes on neutralizing performance and identify 17T2 as a potential candidate for future clinical interventions.

Virus-like particle-mediated delivery of structure-selected neoantigens demonstrates immunogenicity and antitumoral activity in mice

BACKGROUND

Neoantigens are patient- and tumor-specific peptides that arise from somatic mutations. They stand as promising targets for personalized therapeutic cancer vaccines. The identification process for neoantigens has evolved with the use of next-generation sequencing technologies and bioinformatic tools in tumor genomics. However, in-silico strategies for selecting immunogenic neoantigens still have very low accuracy rates, since they mainly focus on predicting peptide binding to Major Histocompatibility Complex (MHC) molecules, which is key but not the sole determinant for immunogenicity. Moreover, the therapeutic potential of neoantigen-based vaccines may be enhanced using an optimal delivery platform that elicits robust de novo immune responses.

METHODS

We developed a novel neoantigen selection pipeline based on existing software combined with a novel prediction method, the Neoantigen Optimization Algorithm (NOAH), which takes into account structural features of the peptide/MHC-I interaction, as well as the interaction between the peptide/MHC-I complex and the TCR, in its prediction strategy. Moreover, to maximize neoantigens' therapeutic potential, neoantigen-based vaccines should be manufactured in an optimal delivery platform that elicits robust de novo immune responses and bypasses central and peripheral tolerance.

RESULTS

We generated a highly immunogenic vaccine platform based on engineered HIV-1 Gag-based Virus-Like Particles (VLPs) expressing a high copy number of each in silico selected neoantigen. We tested different neoantigen-loaded VLPs (neoVLPs) in a B16-F10 melanoma mouse model to evaluate their capability to generate new immunogenic specificities. NeoVLPs were used in in vivo immunogenicity and tumor challenge experiments.

CONCLUSIONS

Our results indicate the relevance of incorporating other immunogenic determinants beyond the binding of neoantigens to MHC-I. Thus, neoVLPs loaded with neoantigens enhancing the interaction with the TCR can promote the generation of de novo antitumor-specific immune responses, resulting in a delay in tumor growth. Vaccination with the neoVLP platform is a robust alternative to current therapeutic vaccine approaches and a promising candidate for future personalized immunotherapy.

Novel Spike-stabilized trimers with improved production protect K18-hACE2 mice and golden Syrian hamsters from the highly pathogenic SARS-CoV-2 Beta variant

Most COVID-19 vaccines are based on the SARS-CoV-2 Spike glycoprotein (S) or their subunits. However, S shows some structural instability that limits its immunogenicity and production, hampering the development of recombinant S-based vaccines. The introduction of the K986P and V987P (S-2P) mutations increases the production and immunogenicity of the recombinant S trimer, suggesting that these two parameters are related. Nevertheless, S-2P still shows some molecular instability and it is produced with low yield. Here we described a novel set of mutations identified by molecular modeling and located in the S2 region of the S-2P that increase its production up to five-fold. Besides their immunogenicity, the efficacy of two representative S-2P-based mutants, S-29 and S-21, protecting from a heterologous SARS-CoV-2 Beta variant challenge was assayed in K18-hACE2 mice (an animal model of severe SARS-CoV-2 disease) and golden Syrian hamsters (GSH) (a moderate disease model). S-21 induced higher level of WH1 and Delta variants neutralizing antibodies than S-2P in K18-hACE2 mice three days after challenge. Viral load in nasal turbinate and oropharyngeal samples were reduced in S-21 and S-29 vaccinated mice. Despite that, only the S-29 protein protected 100% of K18-hACE2 mice from severe disease. When GSH were analyzed, all immunized animals were protected from disease development irrespectively of the immunogen they received. Therefore, the higher yield of S-29, as well as its improved immunogenicity and efficacy protecting from the highly pathogenic SARS-CoV-2 Beta variant, pinpoint the S-29 mutant as an alternative to the S-2P protein for future SARS-CoV-2 vaccine development.

Exploring FeLV-Gag-Based VLPs as a New Vaccine Platform-Analysis of Production and Immunogenicity

Feline leukemia virus (FeLV) is one of the most prevalent infectious diseases in domestic cats. Although different commercial vaccines are available, none of them provides full protection. Thus, efforts to design a more efficient vaccine are needed. Our group has successfully engineered HIV-1 Gag-based VLPs that induce a potent and functional immune response against the HIV-1 transmembrane protein gp41. Here, we propose to use this concept to generate FeLV-Gag-based VLPs as a novel vaccine strategy against this retrovirus. By analogy to our HIV-1 platform, a fragment of the FeLV transmembrane p15E protein was exposed on FeLV-Gag-based VLPs. After optimization of Gag sequences, the immunogenicity of the selected candidates was evaluated in C57BL/6 and BALB/c mice, showing strong cellular and humoral responses to Gag but failing to generate anti-p15E antibodies. Altogether, this study not only tests the versatility of the enveloped VLP-based vaccine platform but also sheds light on FeLV vaccine research.

An engineered HIV-1 Gag-based VLP displaying high antigen density induces strong antibody-dependent functional immune responses

Antigen display on the surface of Virus-Like Particles (VLPs) improves immunogenicity compared to soluble proteins. We hypothesised that immune responses can be further improved by increasing the antigen density on the surface of VLPs. In this work, we report an HIV-1 Gag-based VLP platform engineered to maximise the presence of antigen on the VLP surface. An HIV-1 gp41-derived protein (Min), including the C-terminal part of gp41 and the transmembrane domain, was fused to HIV-1 Gag. This resulted in high-density MinGag-VLPs. These VLPs demonstrated to be highly immunogenic in animal models using either a homologous (VLP) or heterologous (DNA/VLP) vaccination regimen, with the latter yielding 10-fold higher anti-Gag and anti-Min antibody titres. Despite these strong humoral responses, immunisation with MinGag-VLPs did not induce neutralising antibodies. Nevertheless, antibodies were predominantly of an IgG2b/IgG2c profile and could efficiently bind CD16-2. Furthermore, we demonstrated that MinGag-VLP vaccination could mediate a functional effect and halt the progression of a Min-expressing tumour cell line in an in vivo mouse model.

Comparison of Reverse Transcription (RT)-Quantitative PCR and RT-Droplet Digital PCR for Detection of Genomic and Subgenomic SARS-CoV-2 RNA

Most individuals acutely infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) exhibit mild symptoms. However, 10 to 20% of those infected develop long-term symptoms, referred to as post-coronavirus disease 2019 (COVID-19) condition (PCC). One hypothesis is that PCC might be exacerbated by viral persistence in tissue sanctuaries. Therefore, the accurate detection and quantification of SARS-CoV-2 are not only necessary for viral load monitoring but also crucial for detecting long-term viral persistence and determining whether viral replication is occurring in tissue reservoirs. In this study, the sensitivity and robustness of reverse transcription (RT)-droplet digital PCR (ddPCR) and RT-quantitative PCR (qPCR) techniques have been compared for the detection and quantification of SARS-CoV-2 genomic and subgenomic RNAs from oropharyngeal swabs taken from confirmed SARS-CoV-2-positive, SARS-CoV-2-exposed, and nonexposed individuals as well as from samples from mice infected with SARS-CoV-2. Our data demonstrated that both techniques presented equivalent results in the mid- and high-viral-load ranges. Additionally, RT-ddPCR was more sensitive than RT-qPCR in the low-viral-load range, allowing the accurate detection of positive results in individuals exposed to the virus. Overall, these data suggest that RT-ddPCR might be an alternative to RT-qPCR for detecting low viral loads in samples and for assessing viral persistence in samples from individuals with PCC. IMPORTANCE We developed one-step reverse transcription (RT)-droplet digital PCR (ddPCR) protocols to detect SARS-CoV-2 RNA and compared them to the gold-standard RT-quantitative PCR (RT-qPCR) method. RT-ddPCR was more sensitive than RT-qPCR in the low-viral-load range, while both techniques were equivalent in the mid- and high-viral-load ranges. Overall, these results suggest that RT-ddPCR might be a viable alternative to RT-qPCR when it comes to detecting low viral loads in samples, which is a highly relevant issue for determining viral persistence in as-yet-unknown tissue reservoirs in individuals suffering from post-COVID conditions or long COVID.

Remodeling the bladder tumor immune microenvironment by mycobacterial species with changes in their cell envelope composition

Intravesical BCG instillation after bladder tumor resection is the standard treatment for non-muscle invasive bladder cancer; however, it is not always effective and frequently has undesirable side effects. Therefore, new strategies that improve the clinical management of patients are urgently needed. This study aimed to comprehensively evaluate the bladder tumor immune microenvironment profile after intravesical treatment with a panel of mycobacteria with variation in their cell envelope composition and its impact on survival using an orthotopic murine model to identify more effective and safer therapeutic strategies. tumor-bearing mice were intravesically treated with a panel of BCG and M. brumae cultured under different conditions. Untreated tumor-bearing mice and healthy mice were also included as controls. After mycobacterial treatments, the infiltrating immune cell populations in the bladder were analysed by flow cytometry. We provide evidence that mycobacterial treatment triggered a strong immune infiltration into the bladder, with BCG inducing higher global absolute infiltration than M. brumae. The induced global immune microenvironment was strikingly different between the two mycobacterial species, affecting both innate and adaptive immunity. Compared with M. brumae, BCG treated mice exhibited a more robust infiltration of CD4⁺ and CD8⁺ T-cells skewed toward an effector memory phenotype, with higher frequencies of NKT cells, neutrophils/gMDSCs and monocytes, especially the inflammatory subset, and higher CD4⁺ T_EM/CD4⁺ T_reg and CD8⁺ T_EM/CD4⁺ T_reg ratios. Conversely, M. brumae treatment triggered higher proportions of total activated immune cells and activated CD4⁺ and CD8⁺ T_EM cells and lower ratios of CD4⁺ T_EM cells/CD4⁺ T_regs, CD8⁺ T_EM cells/CD4⁺ T_regs and inflammatory/reparative monocytes. Notably, the mycobacterial cell envelope composition in M. brumae had a strong impact on the immune microenvironment, shaping the B and myeloid cell compartment and T-cell maturation profile and thus improving survival. Overall, we demonstrate that the bladder immune microenvironment induced by mycobacterial treatment is species specific and shaped by mycobacterial cell envelope composition. Therefore, the global bladder immune microenvironment can be remodelled, improving the quality of infiltrating immune cells, the balance between inflammatory and regulatory/suppressive responses and increasing survival.

Efficacy of SARS-CoV-2 vaccination in patients with monoclonal gammopathies: A cross sectional study

SARS-CoV-2 vaccination is the most effective strategy to protect individuals with haematologic malignancies against severe COVID-19, while eliciting limited vaccine responses. We characterized the humoral responses following 3 mo after mRNA-based vaccines in individuals at different plasma-cell disease stages: monoclonal gammopathy of undetermined significance (MGUS), smoldering multiple myeloma (SMM), and multiple myeloma on first-line therapy (MM), compared with a healthy population. Plasma samples from uninfected MM patients showed lower SARS-CoV-2-specific antibody levels and neutralization capacity compared with MGUS, SMM, and healthy individuals. Importantly, COVID-19 recovered MM individuals presented significantly higher plasma neutralization capacity compared with their uninfected counterparts, highlighting that hybrid immunity elicit stronger immunity even in this immunocompromised population. No differences in the vaccine-induced humoral responses were observed between uninfected MGUS, SMM and healthy individuals. In conclusion, MGUS and SMM patients could be SARS-CoV-2 vaccinated following the vaccine recommendations for the general population, whereas a tailored monitoring of the vaccine-induced immune responses should be considered in uninfected MM patients.

Heterogeneous Infectivity and Pathogenesis of SARS-CoV-2 Variants Beta, Delta and Omicron in Transgenic K18-hACE2 and Wildtype Mice

The emerging SARS-CoV-2 variants of concern (VOCs) may display enhanced transmissibility, more severity and/or immune evasion; however, the pathogenesis of these new VOCs in experimental SARS-CoV-2 models or the potential infection of other animal species is not completely understood. Here we infected K18-hACE2 transgenic mice with B.1, B.1.351/Beta, B.1.617.2/Delta and BA.1.1/Omicron isolates and demonstrated heterogeneous infectivity and pathogenesis. B.1.351/Beta variant was the most pathogenic, while BA.1.1/Omicron led to lower viral RNA in the absence of major visible clinical signs. In parallel, we infected wildtype (WT) mice and confirmed that, contrary to B.1 and B.1.617.2/Delta, B.1.351/Beta and BA.1.1/Omicron can infect them. Infection in WT mice coursed without major clinical signs and viral RNA was transient and undetectable in the lungs by day 7 post-infection. In silico modeling supported these findings by predicting B.1.351/Beta receptor binding domain (RBD) mutations result in an increased affinity for both human and murine ACE2 receptors, while BA.1/Omicron RBD mutations only show increased affinity for murine ACE2.

The Genome-wide Methylation Profile of CD4+ T Cells From Individuals With Human Immunodeficiency Virus (HIV) Identifies Distinct Patterns Associated With Disease Progression

BACKGROUND

Human genetic variation-mostly in the human leukocyte antigen (HLA) and C-C chemokine receptor type 5 (CCR5) regions-explains 25% of the variability in progression of human immunodeficiency virus (HIV) infection. However, it is also known that viral infections can modify cellular DNA methylation patterns. Therefore, changes in the methylation of cytosine-guanine (CpG) islands might modulate progression of HIV infection.

METHODS

In total, 85 samples were analyzed: 21 elite controllers (EC), 21 subjects with HIV before combination antiretroviral therapy (cART) (viremic, 93 325 human immunodeficiency virus type 1 [HIV-1] RNA copies/mL) and under suppressive cART (cART, median of 17 months, <50 HIV-1 RNA copies/mL), and 22 HIV-negative donors (HIVneg). We analyzed the methylation pattern of 485 577 CpG in DNA from peripheral CD4+ T lymphocytes. We selected the most differentially methylated gene (TNF) and analyzed its specific methylation, messenger RNA (mRNA) expression, and plasma protein levels in 5 individuals before and after initiation of cART.

RESULTS

We observed 129 methylated CpG sites (associated with 43 gene promoters) for which statistically significant differences were recorded in viremic versus HIVneg, 162 CpG sites (55 gene promoters) in viremic versus cART, 441 CpG sites (163 gene promoters) in viremic versus EC, but none in EC versus HIVneg. The TNF promoter region was hypermethylated in viremic versus HIVneg, cART, and EC. Moreover, we observed greater plasma levels of TNF in viremic individuals than in EC, cART, and HIVneg.

CONCLUSIONS

Our study shows that genome methylation patterns vary depending on HIV infection status and progression profile and that these variations might have an impact on controlling HIV infection in the absence of cART.

A Longitudinal Analysis Reveals Early Activation and Late Alterations in B Cells During Primary HIV Infection in Mozambican Adults

Primary HIV infection (PHI) and subsequent chronic infection alter B-cell compartment. However, longitudinal analysis defining the dynamics of B-cell alterations are still limited. We longitudinally studied B-cell subsets in individuals followed for 1 year after PHI (n = 40). Treated and untreated chronic HIV infected (n = 56) and HIV-uninfected individuals (n = 58) were recruited as reference groups at the Manhiça District in Mozambique. B cells were analyzed by multicolor flow-cytometry. Anti-HIV humoral response and plasma cytokines were assessed by ELISA or Luminex-based technology. A generalized activation of B cells induced by HIV occurs early after infection and is characterized by increases in Activated and Tissue-like memory cells, decreases in IgM-IgD- (switched) and IgM-only B cells. These alterations remain mostly stable until chronic infection and are reverted in part by ART. In contrast, other parameters followed particular dynamics: PD-1 expression in memory cells decreases progressively during the first year of infection, Transitional B cells expand at month 3-4 after infection, and Marginal zone-like B cells show a late depletion. Plasmablasts expand 2 months after infection linked to plasma viral load and anti-p24 IgG3 responses. Most of well-defined changes induced by HIV in B-cell activation and memory subsets are readily observed after PHI, lasting until ART initiation. However, subsequent changes occur after sustained viral infection. These data indicate that HIV infection impacts B cells in several waves over time, and highlight that early treatment would result in beneficial effects on the B-cell compartment.

Dissemination of Mycobacterium tuberculosis is associated to a SIGLEC1 null variant that limits antigen exchange via trafficking extracellular vesicles

The identification of individuals with null alleles enables studying how the loss of gene function affects infection. We previously described a non-functional variant in SIGLEC1, which encodes the myeloid-cell receptor Siglec-1/CD169 implicated in HIV-1 cell-to-cell transmission. Here we report a significant association between the SIGLEC1 null variant and extrapulmonary dissemination of Mycobacterium tuberculosis (Mtb) in two clinical cohorts comprising 6,256 individuals. Local spread of bacteria within the lung is apparent in Mtb-infected Siglec-1 knockout mice which, despite having similar bacterial load, developed more extensive lesions compared to wild type mice. We find that Siglec-1 is necessary to induce antigen presentation through extracellular vesicle uptake. We postulate that lack of Siglec-1 delays the onset of protective immunity against Mtb by limiting antigen exchange via extracellular vesicles, allowing for an early local spread of mycobacteria that increases the risk for extrapulmonary dissemination.

Anti-Siglec-1 antibodies block Ebola viral uptake and decrease cytoplasmic viral entry

Several Ebola viruses cause outbreaks of lethal haemorrhagic fever in humans, but developing therapies tackle only Zaire Ebola virus. Dendritic cells (DCs) are targets of this infection in vivo. Here, we found that Ebola virus entry into activated DCs requires the sialic acid-binding Ig-like lectin 1 (Siglec-1/CD169), which recognizes sialylated gangliosides anchored to viral membranes. Blockage of the Siglec-1 receptor by anti-Siglec-1 monoclonal antibodies halted Ebola viral uptake and cytoplasmic entry, offering cross-protection against other ganglioside-containing viruses such as human immunodeficiency virus type 1.

Combined assessment of peritumoral Th1/Th2 polarization and peripheral immunity as a new biomarker in the prediction of BCG response in patients with high-risk NMIBC

Intravesical Bacille Calmette-Guérin (BCG) remains the most effective treatment for high-risk non-muscle-invasive bladder cancer (NMIBC), unfortunately there is no validated biomarker to predict clinical outcome. Here we tried to explore the possibility that a combination of the density of peritumoral infiltrating cells (Th1, Th2 and PD-L1) and the composition of peripheral immune cells (neutrophil and lymphocyte counts) could generate a more reliable prognostic biomarker. Twenty-two patients with high-risk NMIBC treated with BCG (10 BCG nonresponders and 12 BCG responders) were selected. BCG responders had significantly lower level of peritumoral T-bet⁺ cells with an associated higher GATA-3⁺/T-bet⁺ ratio (p = 0.04, p = 0.02, respectively). Furthermore, the immune polarization in tissue (GATA-3⁺/T-bet⁺ ratio) adjusted for the systemic inflammation (neutrophil-to-lymphocyte ratio) showed a significantly higher association with the BCG response (p = 0.004). A survival analysis demonstrated prolonged recurrence-free survival (RFS) in patients with a lower T-bet⁺/Lymphocyte ratio and higher GTR/NLR (p = 0.01). No association was observed between peritumoral PD-L1⁺ expression and the BCG response. In conclusion, alterations in overall immune function, both local and systemic, may influence the therapeutic response to BCG, therefore a combined analysis of tumoral (Th2/Th1 ratio) and peripheral (NLR) immune composition prior to treatment may be a promising approach to predict the BCG response in high-risk NMIBC patients.

Dynamics of CD4 and CD8 T-Cell Subsets and Inflammatory Biomarkers during Early and Chronic HIV Infection in Mozambican Adults

During primary HIV infection (PHI), there is a striking cascade response of inflammatory cytokines and many cells of the immune system show altered frequencies and signs of extensive activation. These changes have been shown to have a relevant role in predicting disease progression; however, the challenges of identifying PHI have resulted in a lack of critical information about the dynamics of early pathogenic events. We studied soluble inflammatory biomarkers and changes in T-cell subsets in individuals at PHI (n = 40), chronic HIV infection (CHI, n = 56), and HIV-uninfected (n = 58) recruited at the Manhiça District Hospital in Mozambique. Plasma levels of 49 biomarkers were determined by Luminex and ELISA. T-cell immunophenotyping was performed by multicolor flow cytometry. Plasma HIV viremia, CD4, and CD8 T cell counts underwent rapid stabilization after PHI. However, several immunological parameters, including Th1-Th17 CD4 T cells and activation or exhaustion of CD8 T cells continued decreasing until more than 9 months postinfection. Importantly, no sign of immunosenescence was observed over the first year of HIV infection. Levels of IP-10, MCP-1, BAFF, sCD14, tumor necrosis factor receptor-2, and TRAIL were significantly overexpressed at the first month of infection and underwent a prompt decrease in the subsequent months while, MIG and CD27 levels began to increase 1 month after infection and remained overexpressed for almost 1 year postinfection. Early levels of soluble biomarkers were significantly associated with subsequently exhausted CD4 T-cells or with CD8 T-cell activation. Despite rapid immune control of virus replication, the stabilization of the T-cell subsets occurs months after viremia and CD4 count plateau, suggesting persistent immune dysfunction and highlighting the potential benefit of early treatment initiation that could limit immunological damage.

An efficient algorithm to perform multiple testing in epistasis screening

BACKGROUND

Research in epistasis or gene-gene interaction detection for human complex traits has grown over the last few years. It has been marked by promising methodological developments, improved translation efforts of statistical epistasis to biological epistasis and attempts to integrate different omics information sources into the epistasis screening to enhance power. The quest for gene-gene interactions poses severe multiple-testing problems. In this context, the maxT algorithm is one technique to control the false-positive rate. However, the memory needed by this algorithm rises linearly with the amount of hypothesis tests. Gene-gene interaction studies will require a memory proportional to the squared number of SNPs. A genome-wide epistasis search would therefore require terabytes of memory. Hence, cache problems are likely to occur, increasing the computation time. In this work we present a new version of maxT, requiring an amount of memory independent from the number of genetic effects to be investigated. This algorithm was implemented in C++ in our epistasis screening software MBMDR-3.0.3. We evaluate the new implementation in terms of memory efficiency and speed using simulated data. The software is illustrated on real-life data for Crohn's disease.

RESULTS

In the case of a binary (affected/unaffected) trait, the parallel workflow of MBMDR-3.0.3 analyzes all gene-gene interactions with a dataset of 100,000 SNPs typed on 1000 individuals within 4 days and 9 hours, using 999 permutations of the trait to assess statistical significance, on a cluster composed of 10 blades, containing each four Quad-Core AMD Opteron(tm) Processor 2352 2.1 GHz. In the case of a continuous trait, a similar run takes 9 days. Our program found 14 SNP-SNP interactions with a multiple-testing corrected p-value of less than 0.05 on real-life Crohn's disease (CD) data.

CONCLUSIONS

Our software is the first implementation of the MB-MDR methodology able to solve large-scale SNP-SNP interactions problems within a few days, without using much memory, while adequately controlling the type I error rates. A new implementation to reach genome-wide epistasis screening is under construction. In the context of Crohn's disease, MBMDR-3.0.3 could identify epistasis involving regions that are well known in the field and could be explained from a biological point of view. This demonstrates the power of our software to find relevant phenotype-genotype higher-order associations.

AUC-RF: a new strategy for genomic profiling with random forest

OBJECTIVE

Genomic profiling, the use of genetic variants at multiple loci simultaneously for the prediction of disease risk, requires the selection of a set of genetic variants that best predicts disease status. The goal of this work was to provide a new selection algorithm for genomic profiling.

METHODS

We propose a new algorithm for genomic profiling based on optimizing the area under the receiver operating characteristic curve (AUC) of the random forest (RF). The proposed strategy implements a backward elimination process based on the initial ranking of variables.

RESULTS AND CONCLUSIONS

We demonstrate the advantage of using the AUC instead of the classification error as a measure of predictive accuracy of RF. In particular, we show that the use of the classification error is especially inappropriate when dealing with unbalanced data sets. The new procedure for variable selection and prediction, namely AUC-RF, is illustrated with data from a bladder cancer study and also with simulated data. The algorithm is publicly available as an R package, named AUCRF, at http://cran.r-project.org/.

Model-based multifactor dimensionality reduction for detecting epistasis in case-control data in the presence of noise

Analyzing the combined effects of genes and/or environmental factors on the development of complex diseases is a great challenge from both the statistical and computational perspective, even using a relatively small number of genetic and nongenetic exposures. Several data-mining methods have been proposed for interaction analysis, among them, the Multifactor Dimensionality Reduction Method (MDR) has proven its utility in a variety of theoretical and practical settings. Model-Based Multifactor Dimensionality Reduction (MB-MDR), a relatively new MDR-based technique that is able to unify the best of both nonparametric and parametric worlds, was developed to address some of the remaining concerns that go along with an MDR analysis. These include the restriction to univariate, dichotomous traits, the absence of flexible ways to adjust for lower order effects and important confounders, and the difficulty in highlighting epistatic effects when too many multilocus genotype cells are pooled into two new genotype groups. We investigate the empirical power of MB-MDR to detect gene-gene interactions in the absence of any noise and in the presence of genotyping error, missing data, phenocopy, and genetic heterogeneity. Power is generally higher for MB-MDR than for MDR, in particular in the presence of genetic heterogeneity, phenocopy, or low minor allele frequencies.

Improving strategies for detecting genetic patterns of disease susceptibility in association studies

The analysis of gene interactions and epistatic patterns of susceptibility is especially important for investigating complex diseases such as cancer characterized by the joint action of several genes. This work is motivated by a case-control study of bladder cancer, aimed at evaluating the role of both genetic and environmental factors in bladder carcinogenesis. In particular, the analysis of the inflammation pathway is of interest, for which information on a total of 282 SNPs in 108 genes involved in the inflammatory response is available. Detecting and interpreting interactions with such a large number of polymorphisms is a great challenge from both the statistical and the computational perspectives. In this paper we propose a two-stage strategy for identifying relevant interactions: (1) the use of a synergy measure among interacting genes and (2) the use of the model-based multifactor dimensionality reduction method (MB-MDR), a model-based version of the MDR method, which allows adjustment for confounders.