Search

International audience

Transformation to aggressive disease histologies generates formidable clinical challenges across cancers, but biological insights remain few. We modeled the genetic heterogeneity of chronic lymphocytic leukemia (CLL) through multiplexed in vivo CRISPR-Cas9 B-cell editing of recurrent CLL loss-of-function drivers in mice and recapitulated the process of transformation from indolent CLL into large cell lymphoma [i.e., Richter syndrome (RS)]. Evolutionary trajectories of 64 mice carrying diverse combinatorial gene assortments revealed coselection of mutations in Trp53, Mga, and Chd2 and the dual impact of clonal Mga/Chd2 mutations on E2F/MYC and interferon signaling dysregulation. Comparative human and murine RS analyses demonstrated tonic PI3K signaling as a key feature of transformed disease, with constitutive activation of the AKT and S6 kinases, downmodulation of the PTEN phosphatase, and convergent activation of MYC/PI3K transcriptional programs underlying enhanced sensitivity to MYC/mTOR/PI3K inhibition. This robust experimental system presents a unique framework to study lymphoid biology and therapy.Significance:Mouse models reflective of the genetic complexity and heterogeneity of human tumors remain few, including those able to recapitulate transformation to aggressive disease histologies. Herein, we model CLL transformation into RS through multiplexed in vivo gene editing, providing key insight into the pathophysiology and therapeutic vulnerabilities of transformed disease.This article is highlighted in the In This Issue feature, p. 101

Transformation to aggressive disease histologies generates formidable clinical challenges across cancers, but biological insights remain few. We modeled the genetic heterogeneity of chronic lymphocytic leukemia (CLL) through multiplexed in vivo CRISPR-Cas9 B-cell editing of recurrent CLL loss-of-function drivers in mice and recapitulated the process of transformation from indolent CLL into large cell lymphoma [i.e., Richter syndrome (RS)]. Evolutionary trajectories of 64 mice carrying diverse combinatorial gene assortments revealed coselection of mutations in Trp53, Mga, and Chd2 and the dual impact of clonal Mga/Chd2 mutations on E2F/MYC and interferon signaling dysregulation. Comparative human and murine RS analyses demonstrated tonic PI3K signaling as a key feature of transformed disease, with constitutive activation of the AKT and S6 kinases, downmodulation of the PTEN phosphatase, and convergent activation of MYC/PI3K transcriptional programs underlying enhanced sensitivity to MYC/mTOR/PI3K inhibition. This robust experimental system presents a unique framework to study lymphoid biology and therapy.Significance:Mouse models reflective of the genetic complexity and heterogeneity of human tumors remain few, including those able to recapitulate transformation to aggressive disease histologies. Herein, we model CLL transformation into RS through multiplexed in vivo gene editing, providing key insight into the pathophysiology and therapeutic vulnerabilities of transformed disease.This article is highlighted in the In This Issue feature, p. 101

Supplementary file contains a PDF version of supplementary figures S1-S7 and associated figure legends.

Supplementary file contains all supplementary tables S1-S16 in Excel format in the order in which they appear in the text.

Idiopathic aplastic anemia (IAA) pathophysiology is dominated by autoreactivity of human leukocyte antigen (HLA)-restricted T-cells against antigens presented by hematopoietic stem and progenitor cells (HSPCs). Expansion of PIGA and HLA class I mutant HSPCs have been linked to immune evasion from T-cell mediated pressures. We hypothesized that, in analogy with anti-tumor immunity, the pathophysiological cascade of immune escape in IAA is initiated by immunoediting pressures and culminates with mechanisms of clonal evolution characterized by hits in immune recognition and response genes. To that end, we studied the genetic and transcriptomic make-up of the antigen presentation complexes in a large cohort of patients with IAA and paroxysmal nocturnal hemoglobinuria (PNH) by using single-cell RNA, high throughput DNA sequencing and single nucleotide polymorphism (SNP)-array platforms. At disease onset HSPCs displayed activation of selected HLA class I and II restricted mechanisms, without extensive inhibition of immune checkpoint apparatus. Using a newly implemented bioinformatic framework we found that not only class I but also class II genes were often impaired by acquisition of genetic aberrations. We also demonstrated the presence of novel somatic alterations in immune genes possibly contributing to the evasion from the autoimmune T-cells. In contrast, these hits were absent in myeloid neoplasia. These aberrations were not mutually exclusive with PNH and did not correlate with the accumulation of myeloid-driver hits. Our findings shed light on the mechanisms of immune activation and escape in IAA and define alternative modes of clonal hematopoiesis.

Background epi-cblC is a recently discovered inherited disorder of intracellular vitamin B12 metabolism associating hematological, neurological, and cardiometabolic outcomes. It is produced by an epimutation at the promoter common to CCDC163P and MMACHC, which results from an aberrant antisense transcription due to splicing mutations in the antisense PRDX1 gene neighboring MMACHC. We studied whether the aberrant transcription produced a second epimutation by encompassing the CpG island of the TESK2 gene neighboring CCDC163P. Methods We unraveled the methylome architecture of the CCDC163P–MMACHC CpG island (CpG:33) and the TESK2 CpG island (CpG:51) of 17 epi-cblC cases. We performed an integrative analysis of the DNA methylome profiling, transcriptome reconstruction of RNA-sequencing (RNA-seq), chromatin immunoprecipitation sequencing (ChIP-Seq) of histone H3, and transcription expression of MMACHC and TESK2. Results The PRDX1 splice mutations and activation of numerous cryptic splice sites produced antisense readthrough transcripts encompassing the bidirectional MMACHC/CCDC163P promoter and the TESK2 promoter, resulting in the silencing of both the MMACHC and TESK2 genes through the deposition of SETD2-dependent H3K36me3 marks and the generation of epimutations in the CpG islands of the two promoters. Conclusions The antisense readthrough transcription of the mutated PRDX1 produces an epigenetic silencing of MMACHC and TESK2. We propose using the term 'epi-digenism' to define this epigenetic disorder that affects two genes. Epi-cblC is an entity that differs from cblC. Indeed, the PRDX1 and TESK2 altered expressions are observed in epi-cblC but not in cblC, suggesting further evaluating the potential consequences on cancer risk and spermatogenesis.

epi-cblC is a recently discovered inherited disorder of intracellular vitamin BWe unraveled the methylome architecture of the CCDC163P-MMACHC CpG island (CpG:33) and the TESK2 CpG island (CpG:51) of 17 epi-cblC cases. We performed an integrative analysis of the DNA methylome profiling, transcriptome reconstruction of RNA-sequencing (RNA-seq), chromatin immunoprecipitation sequencing (ChIP-Seq) of histone H3, and transcription expression of MMACHC and TESK2.The PRDX1 splice mutations and activation of numerous cryptic splice sites produced antisense readthrough transcripts encompassing the bidirectional MMACHC/CCDC163P promoter and the TESK2 promoter, resulting in the silencing of both the MMACHC and TESK2 genes through the deposition of SETD2-dependent H3K36me3 marks and the generation of epimutations in the CpG islands of the two promoters.The antisense readthrough transcription of the mutated PRDX1 produces an epigenetic silencing of MMACHC and TESK2. We propose using the term 'epi-digenism' to define this epigenetic disorder that affects two genes. Epi-cblC is an entity that differs from cblC. Indeed, the PRDX1 and TESK2 altered expressions are observed in epi-cblC but not in cblC, suggesting further evaluating the potential consequences on cancer risk and spermatogenesis.

Background Although radiation therapy represents a core cancer treatment modality, its efficacy is hampered by radioresistance. The effect of ionizing radiations (IRs) is well known regarding their ability to induce genetic alterations; however, their impact on the epigenome landscape in cancer, notably at the CpG dinucleotide resolution, remains to be further deciphered. In addition, no evidence is available regarding the effect of IRs on the DNA methylome profile according to the methionine dependency phenotype, which represents a hallmark of metabolic adaptation in cancer. Methods We used a case–control study design with a fractionated irradiation regimen on four cancerous cell lines representative of HCC (HepG2), melanoma (MeWo and MeWo-LC1, which exhibit opposed methionine dependency phenotypes), and glioblastoma (U251). We performed high-resolution genome-wide DNA methylome profiling using the MethylationEPIC BeadChip on baseline conditions, irradiated cell lines (cumulative dose of 10 Gy), and non-irradiated counterparts. We performed epigenome-wide association studies to assess the effect of IRs and methionine-dependency-oriented analysis by carrying out epigenome-wide conditional logistic regression. We looked for epigenome signatures at the locus and single-probe (CpG dinucleotide) levels and through enrichment analyses of gene ontologies (GO). The EpiMet project was registered under the ID#AAP-BMS_003_211. Results EWASs revealed shared GO annotation pathways associated with increased methylation signatures for several biological processes in response to IRs, including blood circulation, plasma membrane-bounded cell projection organization, cell projection organization, multicellular organismal process, developmental process, and animal organ morphogenesis. Epigenome-wide conditional logistic regression analysis on the methionine dependency phenotype highlighted several epigenome signatures related to cell cycle and division and responses to IR and ultraviolet light. Conclusions IRs generated a variation in the methylation level of a high number of CpG probes with shared biological pathways, including those associated with cell cycle and division, responses to IRs, sustained angiogenesis, tissue invasion, and metastasis. These results provide insight on shared adaptive mechanisms of the epigenome in cancerous cell lines in response to IR. Future experiments should focus on the tryptic association between IRs, the initiation of a radioresistance phenotype, and their interaction with methionine dependency as a hallmark of metabolic adaptation in cancer. Graphical abstract

International audience; Meningiomas are the most common primary tumors of the central nervous system. Based on the 2021 WHO classification, they are classified into three grades reflecting recurrence risk and aggressiveness. However, the WHO’s histopathological criteria defining these grades are somewhat subjective. Together with reliable immunohistochemical proliferation indices, other molecular markers such as those studied with genome-wide epigenetics promise to revamp the current prognostic classification. In this study, 48 meningiomas of various grades were randomly included and explored for DNA methylation with the Infinium MethylationEPIC microarray over 850k CpG sites. We conducted differential and correlative analyses on grade and several proliferation indices and markers, such as mitotic index and Ki-67 or MCM6 immunohistochemistry. We also set up Cox proportional hazard models for extensive associations between CpG methylation and survival. We identified loci highly correlated with cell growth and a targeted methylation signature of regulatory regions persistently associated with proliferation, grade, and survival. Candidate genes under the control of these regions include SMC4, ESRRG, PAX6, DOK7, VAV2, OTX1, and PCDHA-PCDHB-PCDHG, i.e., the protocadherin gene clusters. This study highlights the crucial role played by epigenetic mechanisms in shaping dysregulated cellular proliferation and provides potential biomarkers bearing prognostic and therapeutic value for the clinical management of meningioma.

Context A recent study identified 14 low-frequency coding variants associated with body mass index (BMI) in 718 734 individuals predominantly of European ancestry. Objective We investigated the association of 2 genetic scores (GS) with i) the risk of severe/morbid obesity, ii) BMI variation before weight-loss intervention, iii) BMI change in response to an 18-month lifestyle/behavioral intervention program, and iv) BMI change up to 24 months after bariatric surgery. Methods The 14 low-frequency coding variants were genotyped or sequenced in 342 French adults with severe/morbid obesity and 574 French adult controls from the general population. We built risk and protective GS based on 6 BMI-increasing and 5 BMI-decreasing low-frequency coding variants that were polymorphic in our study. Results While the risk GS was not associated with severe/morbid obesity status, BMI-decreasing low-frequency coding variants were significantly less frequent in patients with severe/morbid obesity than in French adults from the general population. Neither the risk nor the protective GS was associated with BMI before intervention in patients with severe/morbid obesity, nor did they affect BMI change in response to a lifestyle/behavioral modification program. The protective GS was associated with a greater BMI decrease following bariatric surgery. The risk and protective GS were associated with a higher and lower risk of BMI regain after bariatric surgery. Conclusion Our data indicate that in populations of European descent, low-frequency coding variants associated with BMI in the general population also affect the outcomes of bariatric surgery in patients with severe/morbid obesity.

Vitamin BThis study observes two cellular models deficient in vitamin BOur data show that m6A methylation in mRNA could be one of the contributing mechanisms that underlie the neurological manifestations produced by vitamin B

Objective: Deficiency in vitamin B12/folate (methyl donor deficiency [MDD]) produces cardiovascular outcomes during aging and fetal programming effects in newborns of MDD mothers. Whether fetal programming provokes long-term effects on aorta remains largely unknown. Approach and Results: We investigated the impact of fetal programming on ascending aorta of aged rats born from mothers subjected to MDD during gestation/lactation. We performed morphological and molecular examinations of ascending aortas in 21 days- and 400 days-aged rats with initial MDD fetal programming (iMDD) compared with control matched rats. iMDD induces remodeling of the ascending aorta in aged rats, with collagen deposition ( P =0.0008), decreased thickness of elastin ( P P =0.0002). Proteomic analyses, Western blotting, and immunohistochemical examination revealed decreased expression of α-smooth muscle actin, vinculin, SM22α (smooth muscle 22α), and N-cadherin and increased expression of TGF (transforming growth factor) β1. Elastin breaks were correlated to increased neutrophil elastase ( P =0.0002), cathepsin-K ( P =0.0002), cathepsin-S ( P P P =0.02). Proximity Duolink ligation assay showed homocysteinylation of actin-associated and extracellular matrix proteins, including SM22α ( P =0.01), N-cadherin ( P =0.0008), and vinculin ( P =0.001), which was associated with elastin breaks ( P =0.002) and increased expression of MARS (methionyl-tRNA synthetase; involved in irreversible protein homocysteinylation). Furthermore, we observed an inverse relationship between elastin breaks and blood pressure (systolic, P =0.004 and diastolic, P =0.0007). Conclusions: MDD fetal programming produced altered integrity and remodeling of ascending aorta during aging and irreversible MARS-associated homocysteinylation of key proteins of extracellular matrix and elastin homeostasis. This contributes to understanding why homocysteine-lowering vitamin B supplementation fails to relieve vascular complications in adulthood.

International audience

Scope Vitamin B12 deficiency presents various neurological manifestations, such as cognitive dysfunction, mental retardation, or memory impairment. However, the involved molecular mechanisms remain to date unclear. Vitamin B12 is essential for synthesizing S-adenosyl methionine (SAM), the methyl group donor used for almost all transmethylation reactions. Here, we investigated the m6A methylation of mRNAs and their related gene expression in models of vitamin B12 deficiency. Methods and results We studied two cellular models deficient in vitamin B12 and hippocampi of mice knock-out for the CD320 receptor. The decrease in SAM levels resulting from vitamin B12 deficiency was associated with m6 A reduced levels in mRNAs. This was also potentially mediated by the overexpression of the eraser FTO. We further investigated mRNA methylation of some genes involved in neurological functions targeted by the m6A reader YTH proteins. We notably observed a m6A hypermethylation of Prkca mRNA and a consistently increased expression of PKCα, a kinase involved in brain development and neuroplasticity, in the two cellular models. Conclusion Our data show that m6A methylation in mRNA could be one of the contributing mechanisms that underlie the neurological manifestations produced by vitamin B12 deficiency. This article is protected by copyright. All rights reserved.

International audience; Previously, the in vitro growth of cancer stem cells in the form of tumor spheres from five different brain cancer cell lines was found to be methionine-dependent. As this earlier work indicated that ALDH1L2, a folate-dependent mitochondria aldehyde dehydrogenase gene, is upregulated in glioblastoma stem cells, we invalidated this gene using CRISPR-cas 9 technique in this present work. We reported here that this invalidation was effective in U251 glioblastoma cells, and no cas9 off target site could be detected by genome sequencing of the two independent knockout targeting either exon I or exon III. The knockout of ALDH1L2 gene in U251 cells rendered the growth of the cancer stem cells of U251 methionine independent. In addition, a much higher ROS (reactive oxygen radicals) level can be detected in the knockout cells compared to the wild type cells. Our evidence here linked the excessive ROS level of the knockout cells to reduced total cellular NADPH. Our evidence suggested also that the cause of the slower growth of the knockout turmor sphere may be related to its partial differentiation.

Richter syndrome (RS) occurs during the disease course of 2 to 10% of Chronic Lymphocytic Leukemia (CLL). Diffuse Large B-Cell Lymphoma (DLBCL) subtype accounts for 90-95% of RS cases. While presenting with the same morphology as de novo DLBCLs, DLBCL subtype of RS is associated with a very poor outcome. Proteins are the primary cellular biological effectors. Proteome composition is highly dependent on regulatory mechanisms located both upstream and downstream translation (transcriptional regulation, post-translational modifications, protein metabolism). Thus, the analysis of the genome and the transcriptome only allows a putative extrapolation of the expressed proteome. Proteomic studies have been performed in the context of de novo DLBCLs (Fornecker et al. Sci Rep. 2019), unravelling a set of proteins associated with refractoriness. In CLL, it showed different profiles according to IGHV mutational status after B-Cell Receptor activation (Perrot et al. Blood 2011). No proteomic study of RS has been carried out to date. RS sample selection was performed across 7 French institutions affiliated to the FILO (French Innovative Leukemia Organization). A total of 49 fresh frozen biopsies were collected, including 28 de novo DLBCLs and 21 RS, mostly treated with first line R-CHOP. All biopsies were centrally reviewed. RS samples were characterized, with data on CLL-RS clonal relationship and mutational status for a 13-gene panel representing the most frequently mutated genes in CLL. Only DLBCL subtype RS samples with at least 50% tumor purity (range 50-95%) and a minimum 10 mg weight were selected. Peptide measurements were performed using liquid chromatography coupled with tandem mass spectrometry, according to published methods (Muller et al. Sci Rep. 2018). Stringent quality controls were applied to ensure sample integrity, abundance accuracy and overall reproducibility. Proteome reconstruction at peptide and at protein level was achieved with a specifically devised pipeline involving conditional filtering, full normalization, categorization and imputation of missing values. These tools made use of the R/Bioconductor DEP package (Zhang et al. Nat Protoc. 2018). Supervised (Bayesian linear models) and unsupervised (hierarchical clustering, K-means, PCA) analyses were further applied to identify differential protein signatures. These were functionally annotated with ReactomePA (Yu et al. Mol Biosyst. 2016) for pathways and STRING (Szklarczyk et al. Nucleic Acids Res. 2019) for association networks. Extended proteomics analysis identified 1,772 proteins, among which 191 were differentially expressed (False Discovery Rate/FDR < 0.05) in RS samples compared to de novo DLBCLs, with 82 increased and 109 decreased proteins. Hierarchical clustering revealed a highly correlated expression profile of these top candidates and clearly separated the 21 RS and the 28 de novo DLBCL samples (Figure 1). Sample distribution was independent from chemosensitivity/resistance, for DLBCL samples, and also unrelated to GCB/Non-GCB phenotype according to Hans algorithm, Epstein-Barr virus positivity or tumor purity. Functional interactome is an in silico protein-protein interaction network built on published data from the literature and available in public databases. The functional interactome computed from the 82 proteins overexpressed in RS showed a strongly enriched association network (protein-protein interactions; p-value < 1e-16), with an over-representation in BCR pathway, VEGF signaling, JAK-STAT pathway and Interleukin-12, Rho GTPase, and actin coiling (FDR < 0.05; Figure 1). Proteins underexpressed in RS (109) also displayed highly associated interactions (p-value < 1e-16) with a main node including proteins involved in cell death regulation, extracellular matrix organization, regulation of Insulin Growth Factor, and signaling by receptor tyrosine kinase (FDR < 0.05; Figure 1). Here we performed proteomics on a 49-sample cohort of 28 de novo DLBCLs and 21 RS, which revealed a specific and differential signature in RS. This includes increased expression of targets within the druggable signaling pathways BCR and JAK-STAT. Furthermore the decrease in proteins involved in cell death regulation and extracellular matrix organization suggests resistance mechanisms to apoptosis and immune system in RS. Disclosures Dartigeas: Janssen: Honoraria; Roche: Honoraria; Gilead: Other: non-financial support. Tausch:AbbVie: Consultancy, Honoraria, Research Funding; Roche: Consultancy, Honoraria, Research Funding; Janssen-Cilag: Consultancy, Honoraria, Research Funding. Leblond:AstraZeneca: Consultancy, Honoraria; Lilly: Consultancy; Janssen: Consultancy, Honoraria, Research Funding, Speakers Bureau; Amgen: Honoraria; Gilead: Honoraria, Speakers Bureau; AbbVie: Consultancy, Honoraria, Other: Travel/accommodations/expenses, Research Funding, Speakers Bureau; Roche: Consultancy, Honoraria, Other: Travel/accommodations/expenses, Research Funding, Speakers Bureau. Thieblemont:Hospira: Research Funding; Amgen: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Bristol-Myers Squibb: Consultancy, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Cellectis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company), Speakers Bureau; Roche: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company), Research Funding; Celgene: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Incyte: Honoraria; Kite, a Gilead Company: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Novartis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: travel support; AbbVie: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Bayer: Honoraria. Stilgenbauer:Mundipharma: Consultancy, Honoraria, Other, Research Funding; Novartis: Consultancy, Honoraria, Other, Research Funding; Pharmacyclics: Consultancy, Honoraria, Other, Research Funding; Gilead: Consultancy, Honoraria, Other: travel support, Research Funding; Genentech: Consultancy, Honoraria, Other: travel support, Research Funding; Genzyme: Consultancy, Honoraria, Other: travel support, Research Funding; F. Hoffmann-LaRoche: Consultancy, Honoraria, Other: travel support, Research Funding; Celgene: Consultancy, Honoraria, Other: travel support, Research Funding; Boehringer-Ingelheim: Consultancy, Honoraria, Other: travel support, Research Funding; GlaxoSmithKline: Consultancy, Honoraria, Other: travel support, Research Funding; Amgen: Consultancy, Honoraria, Other: travel support, Research Funding; AbbVie: Consultancy, Honoraria, Other: travel support, Research Funding; Janssen-Cilag: Consultancy, Honoraria, Other: travel support, Research Funding. Guieze:abbvie: Honoraria, Other: advisory board, travel funds; janssen cilag: Honoraria, Other: advisory board, travel funds; roche: Other: travle funds; gilead: Honoraria, Other: travel funds; astrazanecka: Honoraria, Other: advisory board. Perrot:Amgen, BMS/Celgene, Janssen, Sanofi, Takeda: Consultancy, Honoraria, Research Funding. Feugier:astrazeneca: Consultancy, Honoraria, Research Funding; abbvie: Consultancy, Honoraria, Research Funding; roche: Consultancy, Honoraria, Research Funding; gilead: Consultancy, Honoraria, Research Funding; janssen: Consultancy, Honoraria, Research Funding. Broséus:AstraZeneca: Consultancy, Honoraria; Janssen: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Gilead: Honoraria.

Richter Syndrome (RS) corresponds to the transformation of Chronic Lymphocytic Leukemia (CLL) or Small Lymphocytic Lymphoma (SLL) into an aggressive lymphoma, in most cases a Diffuse Large B-Cell Lymphoma (DLBCL). RS outcome is variable and still poorly understood. The aim of our study was to analyze the clinical, biological and molecular features liable to predict survival in a retrospective series of newly diagnosed RS from the French Innovative Leukemia Organization (FILO). From 10 French centers, 103 biopsy-confirmed DLBCL subtype RS, diagnosed from 2001 to 2019, were identified. Fresh-frozen biopsies (FB) were available in 58 cases. All biopsies were centrally reviewed. Clinical and biological characteristics at CLL and RS diagnoses including cytogenetics, clonal relationship with the pre-existing CLL, Epstein-Barr virus (EBV) status, cell of origin (COO) according to the Hans algorithm, RS prognostic scores (Tsimberidou et al. J Clin Oncol. 2006, Rossi et al. Blood 2011) as well as treatment and outcomes were collected. Targeted next generation sequencing was used on RS FB for the following gene set: TP53, ATM, SF3B1, NOTCH1, BIRC3, FBXW7, RPS15, EGR2, MYD88, XPO1, POT1, BRAF, and NFKBIE. Overall Survival (OS) was defined as time from RS diagnosis until the date of death or end of follow-up and analyzed using Kaplan-Meier method. Multivariable analysis was performed using Cox regression model for variables with a p-value Clinical and biological characteristics and outcomes were broadly similar between the full 103-patient cohort and the subset of 58 patients with available gene mutation data. The latter only was considered for the subsequent analyses. The median age at CLL/SLL diagnosis was 60 (range 34-81) and 39 patients (67.2%) were male. Prior to RS, 24 (41.38%) received more than 1 line of treatment for CLL. The median time to transformation was 5 years (range 0-22). The median OS from RS diagnosis was 8 months (Figure 1a). The median age at RS diagnosis was 65.5 years (range 42-87). ECOG Performance Status (PS) was >1 in 29/53 patients (54.8%) and 32/56 (57.1%) had a Bulky disease. Elevated Lactate DeHydrogenase (LDH) levels (≥ 1.5N) were found in 38/49 patients (77.6%). Unmutated IGHV was observed in 45/58 (77.6%) RS samples. CLL and RS were clonally related for 29/33 (87.9%) RS with available IGHV sequence at CLL diagnosis. TP53 disruption was detected in 34/56 (60.7%) RS cases including TP53 mutations in 23/58 (39.7%). According to Tsimberidou et al., 10/46 (21.7%) RS were in the low-risk group, 10 in the low-intermediate risk group, 11 (23.9%) in the high-intermediate risk group and the other 15 in the high-risk group. According to the Rossi score, only 4/51 RS (7.8%) were low risk, and 19 (37.3%) and 28 (54.9%) intermediate and high-risk, respectively. The most frequent treatment for RS was R-CHOP-like regimen [38/56 (67.9%)], 5 and 2 patients received autologous or allogeneic stem cell transplantation respectively. Most patients, 40/56 (71.4%) failed to reach complete remission after the first line. By bivariate analysis (Figure 1b-e), ECOG PS, platelet count and TP53 disruption worsened OS (p1, platelet count RS outcome is poor. Here, we focused specifically on Richter cells on diagnostic biopsies for genetic analyses. Unmutated IGHV status was identified as prognostic factor for RS, in addition to the previously described: ECOG PS, platelet count and TP53 disruption. More molecular studies are necessary to increase knowledge about RS and improve the survival of patients with DLBCL-type RS. Disclosures Dartigeas: Janssen: Honoraria; Roche: Honoraria; Gilead: Other: non-financial support. Tausch:Roche: Consultancy, Honoraria, Research Funding; AbbVie: Consultancy, Honoraria, Research Funding; Janssen-Cilag: Consultancy, Honoraria, Research Funding. Leblond:Roche: Consultancy, Honoraria, Other: Travel/accommodations/expenses, Research Funding, Speakers Bureau; AbbVie: Consultancy, Honoraria, Other: Travel/accommodations/expenses, Research Funding, Speakers Bureau; Janssen: Consultancy, Honoraria, Research Funding, Speakers Bureau; Gilead: Honoraria, Speakers Bureau; Amgen: Honoraria; AstraZeneca: Consultancy, Honoraria; Lilly: Consultancy. Thieblemont:Incyte: Honoraria; Bristol-Myers Squibb: Consultancy, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Cellectis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company), Speakers Bureau; AbbVie: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Celgene: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Amgen: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Roche: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company), Research Funding; Novartis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: travel support; Kite, a Gilead Company: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: TRAVEL, ACCOMMODATIONS, EXPENSES (paid by any for-profit health care company); Hospira: Research Funding; Bayer: Honoraria. Laribi:abbvie: Honoraria, Research Funding; amgen: Research Funding; novartis: Honoraria, Research Funding; takeda: Research Funding. Stilgenbauer:F. Hoffmann-LaRoche: Consultancy, Honoraria, Other: travel support, Research Funding; Genentech: Consultancy, Honoraria, Other: travel support, Research Funding; Genzyme: Consultancy, Honoraria, Other: travel support, Research Funding; Gilead: Consultancy, Honoraria, Other: travel support, Research Funding; GlaxoSmithKline: Consultancy, Honoraria, Other: travel support, Research Funding; Janssen-Cilag: Consultancy, Honoraria, Other: travel support, Research Funding; Mundipharma: Consultancy, Honoraria, Other, Research Funding; Novartis: Consultancy, Honoraria, Other, Research Funding; Pharmacyclics: Consultancy, Honoraria, Other, Research Funding; Boehringer-Ingelheim: Consultancy, Honoraria, Other: travel support, Research Funding; Celgene: Consultancy, Honoraria, Other: travel support, Research Funding; Amgen: Consultancy, Honoraria, Other: travel support, Research Funding; AbbVie: Consultancy, Honoraria, Other: travel support, Research Funding. Guieze:abbvie: Honoraria, Other: advisory board, travel funds; janssen cilag: Honoraria, Other: advisory board, travel funds; roche: Other: travle funds; gilead: Honoraria, Other: travel funds; astrazanecka: Honoraria, Other: advisory board. Feugier:gilead: Consultancy, Honoraria, Research Funding; janssen: Consultancy, Honoraria, Research Funding; roche: Consultancy, Honoraria, Research Funding; abbvie: Consultancy, Honoraria, Research Funding; astrazeneca: Consultancy, Honoraria, Research Funding. Broséus:Novartis: Honoraria, Research Funding; Janssen: Honoraria, Research Funding; AstraZeneca: Consultancy, Honoraria; Gilead: Honoraria.

National audience; Le Syndrome de Richter (SR) correspond au développement d’un lymphome secondaire agressif chez les patients présentant une Leucémie Lymphoïde Chronique (LLC). Il possède un phénotype Lymphome B Diffus à Grandes Cellules (LBDGC) dans 90-95 % des cas et se caractérise par une morphologie semblable aux LBDGC primaires avec un phénotype immunohistochimique non-GC dans 90 % des cas. Du fait de sa complexité et de sa chimiorésistance, le pronostic du SR reste sombre. Le protéome est constitué par l’ensemble des protéines. C’est une unité complexe et dynamique qui va varier en fonction des conditions biologiques. Les analyses du domaine protéomique se sont popularisées ces dernières années avec le développement de la spectrométrie de masse en tandem (SDM MS/MS) permettant l’acquisition d’un jeu de données conséquent mais impliquant une multitude de biais (grande hétérogénéité, valeurs manquantes, seuils de détection) à intégrer dans les analyses, avec l’aide de la bioinformatique.Pour la première fois, une étude protéomique a été réalisée par SDM couplée à la chromatographie en phase liquide (LC-MS/MS) en stratégie « bottom-up » de tumeurs ganglionnaires de 21 patients atteints du SR, 9 atteints de Lymphomes B Diffus à Grandes Cellules (LBDGC) primaires chimioréfractaires, 19 LBDGC primaires chimiosensibles et 10 LBDGC secondaires hors SR. Un protocole bioinformatique permettant une reconstruction efficace et une analyse statistique du protéome tant au niveau peptidique que protéique a été mis en place. Cette approche a permis la quantification fine de 1772 protéines et 11385 peptides dans les 59 échantillons. Les analyses statistiques ont mis en lumière 206 protéines et 1437 peptides (correspondant à 271 protéines) différentielles (FDR < 0,05) dans le SR comparativement aux autres LBDGC. L’exploration de ces données a permis de mettre en évidence des protéines impliquées dans la signalisation liée au B-Cell Receptor (BCR), à la mort cellulaire, la matrice extracellulaire et l’inflammation. L’existence de deux profils protéomiques du SR différents est également décrite.Cette étude contribue ainsi à une meilleure compréhension des mécanismes physiopathologiques du SR, permettant l’ouverture de potentielles nouvelles voies thérapeutiques tout en apportant de nouvelles données distinguant le SR des LBDGC primitifs chimioréfractaires et des LBDGC secondaires.

International audience; Vitamin B12 or cobalamin (Cbl) metabolism can be affected by genetic defects leading to defective activity of either methylmalonyl-CoA mutase or methionine synthase or both enzymes. Patients usually present with a wide spectrum of pathologies suggesting that various cellular processes could be affected by modifications in gene expression. We have previously demonstrated that these genetic defects are associated with subcellular mislocalization of RNA binding proteins and subsequent altered nucleo-cytoplasmic shuttling of mRNAs. In order to characterize the possible changes of gene expression in these diseases, we have investigated global gene expression in fibroblasts from patients with cblC and cblG inherited disorders by RNA-seq. The most differentially expressed genes are strongly associated with developmental processes, neurological, ophthalmologic and cardiovascular diseases. These associations are consistent with the clinical presentation of cblC and cblG disorders. Multivariate analysis of transcript processing revealed splicing alterations that led to dramatic changes in cytoskeleton organization, response to stress, methylation of macromolecules and RNA binding. The RNA motifs associated with this differential splicing reflected a potential role of RNA binding proteins such as HuR and HNRNPL. Proteomic analysis confirmed that mRNA processing was significantly disturbed. This study reports a dramatic alteration of gene expression in fibroblasts of patients with cblC and cblG disorders, which resulted partly from disturbed function of RNA binding proteins. These data suggest to evaluate the rescue of the mislocalization of RNA binding proteins as a potential strategy in the treatment of severe cases who are resistant to classical treatments with co-enzyme supplements.

International audience; Dopamine (DA) is a ubiquitous neurotransmitter exerting a range of pleiotropic actions through two DA receptor families, the D1 and the D2. To date in vertebrates, a maximum of four receptor subtypes have been identified within the D1 family, D1 (former D1A), D5 (former D1B), D6 (former D1C and D1D) and D7 (former D1E), while the D2 family encloses five subtypes, D2, D3, D4, D8 (former D2like or D2l) and D9 (former D4-related sequence or D4-rs). In teleosts, no study has investigated in parallel all the DA receptors to identify and localize the whole receptor repertoire from both families. In pikeperch, Sander lucioperca, a species of interest for aquaculture development, the existence, number and location of the DA receptors are totally unknown. To address these questions, RNA-seq with de novo transcriptome reconstruction, functional annotation and phylogenetic analysis were performed to characterize the transcript repertoire of DA receptors in the brain of female pikeperch at the pre-ovulatory period. Ten different cDNA were identified and showed to belong to the D1 family: two D1, one D5a, one D6a and one D6b and to the D2 family: two spliced variants of D2, one D3, one D8 and one D9. Unlike zebrafish, the subtypes D4 and D7 have not yet been isolated in pikeperch. As expected D1, D3, D8 and D9 are mostly expressed in brain parts except for the cerebellum (D1 and D3). The inter-species differences in the number of DA receptors and the inter-organ differences in the gene expression of all receptors support the complexity of the dopaminergic actions in vertebrate.

The molecular mechanisms that underlie the neurological manifestations of patients with inherited diseases of vitamin B12 (cobalamin) metabolism remain to date obscure. We observed transcriptomic changes of genes involved in RNA metabolism and endoplasmic reticulum stress in a neuronal cell model with impaired cobalamin metabolism. These changes were related to the subcellular mislocalization of several RNA binding proteins, including the ELAVL1/HuR protein implicated in neuronal stress, in this cell model and in patient fibroblasts with inborn errors of cobalamin metabolism and Cd320 knockout mice. The decreased interaction of ELAVL1/HuR with the CRM1/exportin protein of the nuclear pore complex and its subsequent mislocalization resulted from hypomethylation at R-217 produced by decreased S-adenosylmethionine and protein methyl transferase CARM1 and dephosphorylation at S221 by increased protein phosphatase PP2A. The mislocalization of ELAVL1/HuR triggered the decreased expression of SIRT1 deacetylase and genes involved in brain development, neuroplasticity, myelin formation, and brain aging. The mislocalization was reversible upon treatment with siPpp2ca, cobalamin, S-adenosylmethionine, or PP2A inhibitor okadaic acid. In conclusion, our data highlight the key role of the disruption of ELAVL1/HuR nuclear export, with genomic changes consistent with the effects of inborn errors of Cbl metabolisms on brain development, neuroplasticity and myelin formation.

HuR regulates cytoplasmic mRNA stability and translatability, and the HuR expression level has been shown to correlate with poor disease outcome in several cancer types; however, the prognostic value and potential pro-oncogenic properties of HuR in meningioma remain unclear. Thus, in the present study, we analysed 85 meningioma tissue samples to establish the relationship between HuR expression, tumour cell proliferation, and/or patient survival. In addition, we examined the anti-proliferative effects of HuR knockdown in two meningioma cell lines (IOMM-Lee and Ben-Men-1) and conducted transcriptome-wide analyses (IOMM-Lee cells) to elucidate the molecular consequences of HuR knockdown. The results of the present study showed HuR cytoplasmic expression to correlate positively with tumour grade (p = 1.2 × 10-8 ) and negatively with progression-free and overall survival (p = 0.01) time in human meningioma tissues. In vitro, siHuR-induced HuR knockdown was shown to reduce the growth of both Ben-Men-1 (p = 2 × 10-8 ) and IOMM-Lee (p = 4 × 10-9 ) cells. Transcriptome analyses revealed HuR knockdown in IOMM-Lee cells to deregulate the HIF1A signalling pathway (p = 1.5 × 10-6 ) and to up-regulate the expression of genes essential for the assembly of the cytoplasmic mRNA processing body, global genome nucleotide-excision repair, poly(A)-specific ribonuclease activity, the positive regulation of apoptosis and of cell cycle arrest, and the negative regulation of RNA splicing [p(FDR)

International audience; Anaplastic oligodendroglioma (AO), IDH‐mutant and 1p/19q codeleted (IDHmut+/1p19qcodel), is a high‐grade glioma with only limited prognostic markers. The primary objective of this study was to evaluate, by immunohistochemistry, the prognostic value of two proliferation markers, MCM6 and Ki‐67, in a large series of IDHmut+/1p19qcodel AO included in the POLA (“Prise en charge des Oligodendrogliomes Anaplasiques”) French national multicenter network. We additionally examined the transcriptome obtained from this series to understand the functional pathways dysregulated with the mRNA overexpression of these two markers. The labelling indices (LI) of MCM6 and Ki‐67 were obtained via computer‐assisted color image analyses on immunostained AO tissues of the cohort (n=220). Furthermore, a subgroup of AO (n=68/220) was used to perform transcriptomic analyses. A high LI of either MCM6 (≥50%) or Ki‐67 (≥ 15%) correlated with shorter overall survival, both in univariate (P=0.013 and P=0.004, respectively) and multivariate analyses (P=0.027; multivariate Cox model including age, mitotic index, MCM6 and Ki‐67). MCM6 and Ki‐67 LI also correlated with overall survival in an additional retrospective cohort of 30 grade II IDHmut+/1p19qcodel oligodendrogliomas. The prognostic value of MCM6 mRNA level was confirmed in The Cancer Genome Atlas (TCGA) IDHmut+/1p19qcodel gliomas. The transcriptomic approach revealed that high transcriptional expressions of MCM6 and MKI67 were both linked positively with cell cycle progression, DNA replication, mitosis, pro‐neural phenotype as well as neurogenesis, and negatively with microglial cell activation, immune response, positive regulation of myelination, oligodendrocyte development, beta‐amyloid binding, and postsynaptic specialization.In conclusion, the overexpression of MCM6 and/or Ki‐67 is independently associated to shorter overall survival in IDHmut+/1p19qcodel AO. These two easy‐to‐use and cost‐effective markers could thus be used concurrently in routine pathology practice. Additionally, the transcriptomic analyses showed that AO with high proliferation index have down‐regulated immune response and lower microglial cells activation, and bears pro‐neural phenotype.

Gestational methyl donor (especially B9 and B12 vitamins) deficiency is involved in birth defects and brain development retardation. The underlying molecular mechanisms that are dysregulated still remain poorly understood, in particular in the cerebellum. As evidenced from previous data, females are more affected than males. In this study, we therefore took advantage of a validated rat nutritional model and performed a microarray analysis on female progeny cerebellum, in order to identify which genes and molecular pathways were disrupted in response to methyl donor deficiency. We found that cerebellum development is altered in female pups, with a decrease of the granular cell layer thickness at postnatal day 21. Furthermore, we investigated the involvement of the Wnt signaling pathway, a major molecular pathway involved in neuronal development and later on in synaptic assembly and neurotransmission processes. We found that Wnt canonical pathway was disrupted following early methyl donor deficiency and that neuronal targets were selectively enriched in the downregulated genes. These results could explain the structural brain defects previously observed and highlighted new genes and a new molecular pathway affected by nutritional methyl donor deprivation.

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.

International audience; Scope: Our study aims to investigate molecular events associated to methyl donor deficiency (MDD) by analyzing the transcriptome and the methylome of MDD rats in liver. Methods and results: Twenty-one-day-old rats born to mothers fed either with a standard diet or a MDD diet during gestation and lactation were compared. From a total of 44 000 probes for 26 456 genes, we found two gene clusters in MDD rats whose expression levels had significant differences compared with controls: 3269 overexpressed (p < 0.0009) and 2841 underexpressed (p < 0.0004) genes. Modifications of DNA methylation were found in the promoter regions of 1032 genes out of 14 981 genes. Ontological analyses revealed that these genes are mainly involved in glucose and lipid metabolism, nervous system, coagulation, ER stress, and mitochondrial function. Conclusion: Putative master genes exhibiting changes in both gene expression and DNA methylation are limited to 266 genes and are mainly involved in the renin-angiotensin system (n = 3), mitochondrion metabolism (n = 18), and phospholipid homeostasis (n = 3). Most of these master genes participate in nonalcoholic fatty liver disease. The adverse effects of MDD on the metabolic process indicate the beneficial impact of folate and vitamin B12, especially during the perinatal period.

International audience; We present a new data mining method based on stochastic analysis (HMM for Hidden Markov Model) and combinatorial methods for discovering new transcriptional factors in bacterial genome sequences. Sigma factor binding sites (SFBSs) were described as patterns of box1 - spacer - box2 corresponding to the -35 and -10 DNA motifs of bacterial promoters. We used a high-order Hidden Markov Model in which the hidden process is a second-order Markov chain. Applied on the genome of the model bacterium Streptomyces coelicolor (2), the a posteriori state probabilities revealed local maxima or peaks whose distribution was enriched in the intergenic sequences (``iPeaks'' for intergenic peaks). Short DNA sequences underlying the iPeaks were extracted and clustered by a hierarchical classification algorithm based on the SmithWaterman local similarity. Some selected motif consensuses were used as box1 (-35 motif) in the search of a potential neighbouring box2 (-10 motif) using a word enumeration algorithm. This new SFBS mining methodology applied on Streptomyces coelicolor was successful to retrieve already known SFBSs and to suggest new potential transcriptional factor binding sites (TFBSs). The well defined SigR regulon (oxidative stress response) was also used as a test quorum to compare first and second-order HMM. Our approach also allowed the preliminary detection of known SFBSs in Bacillus subtilis.

International audience; The Streptococcus genus contains both pathogenic bacteria and bacteria used in the food-processing industry. We are developing a statistical segmentation method to identify heterogeneous sequences such as sequences acquired from recent horizontal transfer or genes weakly or strongly expressed. The method is based on second order Hidden Markov Models (HMM2). After an automatic unsupervised training, this method allows to demarcating some particular areas into a genome. After checking the efficiency of such models on various controls and on chimeric sequences generated in silico, we choose a HMM2 (3-mer, 5 states) to analyse the complete genome sequence of S. Thermophilus CNRZ1066 (1.8 Mb). More the 80 atypical segments were extracted and are currently analysed further.

Poster de présentation de l'ACI FouDanga aux journées JOBIM 2005 (Lyon) : rapport d'avancement à 1 an.; L'accumulation des séquences issues des projets de séquençage oblige la mise en œuvre de méthodes de fouilles de données pour comprendre les mécanismes impliqués dans l'expression, la transmission et l'évolution des gènes. Nous nous intéressons aux méthodes combinatoires et stochastiques permettant de prédire les séquences promotrices et autres petites séquences régulatrices chez les bactéries. Deux approches informatiques sont développées. La première correspond à l'utilisation d'algorithmes de recherche de mots puis de couples de mots sur-représentés dans les régions en amont de gènes orthologues d'espèces phylogénétiquement proches. La seconde correspond à une méthode de fouille de données génomiques sans a priori pour faire émerger des sous-séquences d'ADN dans les régions intergéniques. Le processus de fouille de données se traduit par la spécification de modèles de Markov cachés du second-ordre (HMM2), leur apprentissage et leur utilisation pour faire apparaître des irrégularités dans des grandes séquences d'ADN.

in conjonction with the french national conference on Bioinformatics (JOBIM 2003). Colloque avec actes sans comité de lecture. internationale.; International audience; We have developed a data mining study based on second-order Hidden Markov Models to detect short DNA motifs that appear abnormally frequently at non-random locations in the streptomyces coelicolor genome. Our method has permitted to detect 23 of 30 SigR target promoter sequences in this genome. || Nous décrivons une étude de fouille de données à l'aide de modèles de Markov cachés du second ordre dans laquelle nous détectons des petits motifs d'ADN dans la bactérie S. Coelicolor. Notre méthode a permis de détecter 23 des 30 promoteurs sigR de ce gén

Colloque avec actes et comité de lecture. nationale.; National audience; Nous présentons une méthode de fouille de données génomiques dans laquelle l'utilisateur analyse un signal élaboré pour la circonstance par un HMM d'ordre deux. Ce signal, qui représente une probabilité a posteriori de classer un résidu ou un groupe de résidus nucléotidiques dans un état, permet la localisation de répétitions dans la séquence d'un génome bactérien complet.

Colloque avec actes et comité de lecture. internationale.; International audience; We present a genomic data mining method in which the user describes a signal worked out by a second order HMM. This signal representing the probability to classify a nucleotidic residue or a group of residues in a particular state, allows the localization of repetitions in a complete bacterial genomic sequence.

Stage de DEA. Rapport de stage.; Les modèles de Markov cachés (HMM), introduits dans les années 1960, ont connu un succès important en reconnaissance de la parole où ils se sont imposés comme un des modèles de référence. Leur adaptation à l'analyse de grands fragments nucléotidiques est désormais rendue possible par l'accroissement du nombre de génomes bactériens dont la séquence complète est disponible. Nous avons développé une panoplie d'outils permettant d'une part la segmentation de séquences par des HMM d'ordre deux , et d'autre part la visualisation graphique de cette segmentation. Une analyse fouillée des cosmides issus du séquençage du génome de Streptomyces coelicolor a ainsi pu être entreprise. Ce travail a permis de démontrer la capacité de tels modèles à détecter certains types de motifs réitérés présents sur le chromosome.l'étude détaillée de ces motifs montre que certains d'entre eux pourraient être la cause ou la conséquence des phénomènes d'instabilité génétique chez cette bactérie. || Hidden Markov Models (HMM) are still a reference in speech recognition processing since their introduction in the 1960's. Completely sequenced bacterial genomes provides the datas needed for their analysis by these models. We developed a bunch of tools fo

Colloque avec actes et comité de lecture. nationale.; National audience; Nous développons des méthodes de fouille de données basées sur l'utilisation de modèles Markoviens du second ordre adaptés à l'étude des génomes. Ceux-ci réalisent une segmentation pouvant être observée sous la forme d'un signal stochastique traduisant l'organisation et la structure des motifs d'ADN sous-jacents. Aucune hypothèse 'a priori' n'est effectuée sur le contenu génétique des séquences étudiées. La modélisation du corpus de séquences est réalisée par une étape d'apprentissage automatique qui produit une classification non supervisée des segments nucléotidiques observés sur les différents états des HMM. Une première étape d'apprentissage sur les séquences chromosomiques complètes des bactéries actinomycètes Streptomyces coelicolor, S. avermitilis et Mycobacterium tuberculosis permet l'obtention de trois classes de HMM décrivant chacune un génome. Lors du processus de segmentation, certaines chaînes d'états cachés décrivent des fragments génomiques comme les gènes et les séquences intergéniques alors qu'une autre chaîne se spécialise sur la distribution de motifs d'ADN locaux particuliers. Ceux-ci correspondent à des mots de 5 à 12 nucléotides présents à des fréquences inhabituelles dans les régions intergéniques. Chez S. coelicolor, la classification de 2500 de ces motifs, issus d'une extraction automatique et identifiés dans 1,2 Mb d'ADN génomique, indique que 7% correspondraient à des sites de fixation de facteurs sigma connus (SigR, SigB, WhiG, HrdB) et 5% à des sites de fixation du ribosome ou des terminateurs de transcription potentiels. Concernant le régulon SigR/SigH (réponse au stress oxydant chez les Streptomyces/M. tuberculosis), la mise en oeuvre de cette approche a permis de détecter tous les promoteurs déjà déterminés biologiquement. Enfin, certains de ces motifs ne peuvent être corrélés à des rôles biologiques connus ou prédits à ce jour. Leur classification pourrait mettre en évidence des groupes à propriétés communes et viserait à définir des motifs promoteurs, puis, à terme, des réseaux de gènes co-régulés.