Large-scale Genome-wide Analysis Research Articles

Bifidobacteriaceae diversity in the human microbiome from a large-scale genome-wide analysis

We performed a large-scale genome-wide analysis aiming to investigate the prevalence and strain-level diversity of Bifidobacteriaceae species in the human microbiome. We considered 9,528 publicly available human metagenomes and integrated them with 1,192 isolate genomes from different sources. The prevalence and abundance of Bifidobacteriaceae species in humans was linked to multiple host characteristics: they were reduced in older people and enriched in populations characterized by Westernized lifestyles with geography-specific patterns. Phylogenetic analysis highlighted 110 Bifidobacteriaceae species-level genome bins (SGBs), with 32 found in humans and 8 in food and probiotic sources. Functional annotation revealed a great diversity in carbohydrate-active enzyme families across these SGBs. We found potential subspecies for most of the SGBs prevalent in humans and identified patterns driven by age and geography. We provided evidence that strains used in probiotics were rarely identified in humans, with the only exception represented by Bifidobacterium animalis. We finally evaluated that the abundance of Bifidobacteriaceae species exhibited moderate and variable capabilities to predict health status in case-control studies.

Large-scale genome-wide SNP analysis reveals the rugged (and ragged) landscape of global ancestry, phylogeny, and demographic history in chicken breeds.

The worldwide chicken gene pool encompasses a remarkable, but shrinking, number of divergently selected breeds of diverse origin. This study was a large-scale genome-wide analysis of the landscape of the complex molecular architecture, genetic variability, and detailed structure among 49 populations. These populations represent a significant sample of the world's chicken breeds from Europe (Russia, Czech Republic, France, Spain, UK, etc.), Asia (China), North America (USA), and Oceania (Australia). Based on the results of breed genotyping using the Illumina 60K single nucleotide polymorphism (SNP) chip, a bioinformatic analysis was carried out. This included the calculation of heterozygosity/homozygosity statistics, inbreeding coefficients, and effective population size. It also included assessment of linkage disequilibrium and construction of phylogenetic trees. Using multidimensional scaling, principal component analysis, and ADMIXTURE-assisted global ancestry analysis, we explored the genetic structure of populations and subpopulations in each breed. An overall 49-population phylogeny analysis was also performed, and a refined evolutionary model of chicken breed formation was proposed, which included egg, meat, dual-purpose types, and ambiguous breeds. Such a large-scale survey of genetic resources in poultry farming using modern genomic methods is of great interest both from the viewpoint of a general understanding of the genetics of the domestic chicken and for the further development of genomic technologies and approaches in poultry breeding. In general, whole genome SNP genotyping of promising chicken breeds from the worldwide gene pool will promote the further development of modern genomic science as applied to poultry.

STS-BN: An efficient Bayesian network method for detecting causal SNPs.

Background: The identification of the causal SNPs of complex diseases in large-scale genome-wide association analysis is beneficial to the studies of pathogenesis, prevention, diagnosis and treatment of these diseases. However, existing applicable methods for large-scale data suffer from low accuracy. Developing powerful and accurate methods for detecting SNPs associated with complex diseases is highly desired. Results: We propose a score-based two-stage Bayesian network method to identify causal SNPs of complex diseases for case-control designs. This method combines the ideas of constraint-based methods and score-and-search methods to learn the structure of the disease-centered local Bayesian network. Simulation experiments are conducted to compare this new algorithm with several common methods that can achieve the same function. The results show that our method improves the accuracy and stability compared to several common methods. Our method based on Bayesian network theory results in lower false-positive rates when all correct loci are detected. Besides, real-world data application suggests that our algorithm has good performance when handling genome-wide association data. Conclusion: The proposed method is designed to identify the SNPs related to complex diseases, and is more accurate than other methods which can also be adapted to large-scale genome-wide analysis studies data.

Hunt for mammalian sleep-regulating genes

Genetics is one of the various approaches adopted to understand and control mammalian sleep. Reverse genetics, which is usually applied to analyze sleep in gene-deficient mice, has been the mainstream field of genetic studies on sleep for the past three decades and has revealed that various molecules, including orexin, are involved in sleep regulation. Recently, forward genetic studies in humans and mice have identified gene mutations responsible for heritable sleep abnormalities, such as SIK3, NALCN, DEC2, the neuropeptide S receptor, and β1 adrenergic receptor. Furthermore, the protein kinase A-SIK3 pathway was shown to represent the intracellular neural signaling for sleep need. Large-scale genome-wide analyses of human sleep have been conducted, and many gene loci associated with individual differences in sleep have been found. The development of genome-editing technology and gene transfer by an adeno-associated virus has updated and expanded the genetic studies on mammals. These efforts are expected to elucidate the mechanisms of sleep–wake regulation and develop new therapeutic interventions for sleep disorders.

Genome-wide analyses of human noroviruses provide insights on evolutionary dynamics and evidence of coexisting viral populations evolving under recombination constraints.

Norovirus is a major cause of acute gastroenteritis worldwide. Over 30 different genotypes, mostly from genogroup I (GI) and II (GII), have been shown to infect humans. Despite three decades of genome sequencing, our understanding of the role of genomic diversification across continents and time is incomplete. To close the spatiotemporal gap of genomic information of human noroviruses, we conducted a large-scale genome-wide analyses that included the nearly full-length sequencing of 281 archival viruses circulating since the 1970s in over 10 countries from four continents, with a major emphasis on norovirus genotypes that are currently underrepresented in public genome databases. We provided new genome information for 24 distinct genotypes, including the oldest genome information from 12 norovirus genotypes. Analyses of this new genomic information, together with those publicly available, showed that (i) noroviruses evolve at similar rates across genomic regions and genotypes; (ii) emerging viruses evolved from transiently-circulating intermediate viruses; (iii) diversifying selection on the VP1 protein was recorded in genotypes with multiple variants; (iv) non-structural proteins showed a similar branching on their phylogenetic trees; and (v) contrary to the current understanding, there are restrictions on the ability to recombine different genomic regions, which results in co-circulating populations of viruses evolving independently in human communities. This study provides a comprehensive genetic analysis of diverse norovirus genotypes and the role of non-structural proteins on viral diversification, shedding new light on the mechanisms of norovirus evolution and transmission.

Level and change in bone microarchitectural parameters and their relationship with previous fracture and established bone mineral density loci

Osteoporosis is characterised by a reduction of bone mineral density (BMD) and predisposition to fracture. Bone microarchitecture, measured by high resolution peripheral quantitative computed tomography (HR-pQCT), has been related to fragility fractures and BMD and has been the subject of large-scale genome-wide analysis. We investigated whether fracture was related to baseline values and longitudinal changes in bone microarchitecture and whether bone microarchitecture was associated with established BMD loci. 115 males and 99 females (aged 72-81 at baseline) from the Hertfordshire Cohort Study (HCS) were analysed. Fracture history was determined in 2011-2012 by self-report and vertebral fracture assessment. Participants underwent HR-pQCT scans of the distal radius and tibia in 2011-2012 and 2017. Previous fracture in relation to baseline values and changes in tibial HR-pQCT parameters was examined using sex-adjusted logistic regression with and without adjustment for age, sociodemographic, lifestyle and clinical characteristics; baseline values and changes in parameters associated with previous fracture were then examined in relation to four established BMD loci after adjustment for sex and age. Previous fracture was related to: higher trabecular area (fully-adjusted odds ratio [95% CI] per SD greater baseline value: 2.18 [1.27,3.73], p=0.005); lower total volumetric BMD (0.53 [0.34,0.84], p=0.007), cortical area (0.53 [0.30,0.95], p=0.032), cortical BMD (0.56 [0.36,0.88], p=0.011) and cortical thickness (0.45 [0.27,0.77], p=0.004); and greater declines in trabecular BMD (p=0.001). Associations were robust in sex- and fully-adjusted analysis. Relationships between BMD loci and these HR-pQCT parameters were weak: rs3801387 (WNT16) was related to decline in trabecular BMD (p=0.011) but no other associations were significant (p>0.05). Baseline values of HR-pQCT parameters and greater decline in trabecular BMD were associated with fracture. Change in trabecular BMD was associated with WNT16 which has been demonstrated to influence bone health in murine models and human genome-wide association studies (GWAS).

Large-scale genome-wide analysis links lactic acid bacteria from food with the gut microbiome

Lactic acid bacteria (LAB) are fundamental in the production of fermented foods and several strains are regarded as probiotics. Large quantities of live LAB are consumed within fermented foods, but it is not yet known to what extent the LAB we ingest become members of the gut microbiome. By analysis of 9445 metagenomes from human samples, we demonstrate that the prevalence and abundance of LAB species in stool samples is generally low and linked to age, lifestyle, and geography, with Streptococcus thermophilus and Lactococcus lactis being most prevalent. Moreover, we identify genome-based differences between food and gut microbes by considering 666 metagenome-assembled genomes (MAGs) newly reconstructed from fermented food microbiomes along with 154,723 human MAGs and 193,078 reference genomes. Our large-scale genome-wide analysis demonstrates that closely related LAB strains occur in both food and gut environments and provides unprecedented evidence that fermented foods can be indeed regarded as a possible source of LAB for the gut microbiome.

Risk factors for valvular calcification.

Recent literature is examined to identify established and emerging risk factors for valvular calcification, specifically calcific aortic valve disease and mitral annular calcification. Strong evidence implicates older age, male sex, cigarette smoking, elevated blood pressure, dyslipidaemia, adiposity, and mineral metabolism as risk factors for calcific aortic valve disease. Emerging evidence suggests family history and lipoprotein(a) are additional risk factors. Recently, large-scale genome-wide analyses have identified robust associations for LPA, PALMD, and TEX41 with aortic stenosis. Factors predisposing to mitral annular calcification are less well characterized. Older age, cigarette smoking, increased BMI, kidney dysfunction, and elevated triglycerides are associated with greater risk of mitral annular calcification, but conflicting evidence exists for sex and C-reactive protein. Established and emerging risk factors for calcific aortic valve disease, including some that overlap with atherosclerosis, may represent targets for pharmacological intervention. Mitral annular calcification is comparatively less well understood though some atherosclerosis risk factors do appear to increase risk.

Genome-wide copy number analyses of samples from LACE-Bio project identify novel prognostic and predictive markers in early stage non-small cell lung cancer.

Adjuvant chemotherapy (ACT) provides modest benefit in resected non-small cell lung cancer (NSCLC) patients. Genome-wide studies have identified gene copy number aberrations (CNA), but their prognostic implication is unknown. DNA from 1,013 FFPE tumor samples from three pivotal multicenter randomized trials (ACT vs. control) in the LACE-Bio consortium (median follow-up: 5.2 years) was successfully extracted, profiled using a molecular inversion probe SNP assay, normalized relative to a pool of normal tissues and segmented. Minimally recurrent regions were identified. P values were adjusted to control the false discovery rate (Q values). A total of 976 samples successfully profiled, 414 (42%) adenocarcinoma (ADC), 430 (44%) squamous cell carcinoma (SCC) and 132 (14%) other NSCLC; 710 (73%) males. We identified 431 recurrent regions, with on average 51 gains and 43 losses; 253 regions (59%) were ≤3 Mb. Most frequent gains (up to 48%) were on chr1, 3q, 5p, 6p, 8q, 22q; most frequent losses (up to 40%) on chr3p, 8p, 9p. CNA frequency of 195 regions was significantly different (Q≤0.05) between ADC and SCC. Fourteen regions (7p11-12, 9p21, 18q12, and 19p11-13) were associated with disease-free survival (DFS) (univariate P≤0.005, Q<0.142), with poorer DFS for losses of regions including CDKN2A/B [hazard ratio (HR) for 2-fold lower CN: 1.5 (95% CI: 1.2-1.9), P<0.001, Q=0.020] and STK11 [HR =2.4 (1.3-4.3), P=0.005, Q=0.15]. Chromosomal instability was associated with poorer DFS (HR =1.5, P=0.015), OS (HR =1.2, P=0.189) and lung-cancer specific survival (HR =1.7, P=0.003). These large-scale genome-wide analyses of gene CNA provide new candidate prognostic markers for stage I-III NSCLC.

Genome-wide analysis of ionotropic receptor gene repertoire in Lepidoptera with an emphasis on its functions of Helicoverpa armigera

The functions of the Ionotropic Receptor (IR) family have been well studied in Drosophila melanogaster, but only limited information is available in Lepidoptera. Here, we conducted a large-scale genome-wide analysis of the IR gene repertoire in 13 moths and 16 butterflies. Combining a homology-based approach and manual efforts, totally 996 IR candidates are identified including 31 pseudogenes and 825 full-length sequences, representing the most current comprehensive annotation in lepidopteran species. The phylogeny, expression and sequence characteristics classify Lepidoptera IRs into three sub-families: antennal IRs (A-IRs), divergent IRs (D-IRs) and Lepidoptera-specific IRs (LS-IRs), which is distinct from the case of Drosophila IRs. In comparison to LS-IRs and D-IRs, A-IRs members share a higher degree of protein identity and are distinguished into 16 orthologous groups in the phylogeny, showing conservation of gene structure. Analysis of selective forces on 27 orthologous groups reveals that these lepidopteran IRs have evolved under strong purifying selection (dN/dS≪1). Most notably, lineage-specific gene duplications that contribute primarily to gene number variations across Lepidoptera not only exist in D-IRs, but are present in the two other sub-families including members of IR41a, 76b, 87a, 100a and 100b. Expression profiling analysis reveals that over 80% (21/26) of Helicoverpa armigera A-IRs are expressed more highly in antennae of adults or larvae than other tissues, consistent with its proposed function in olfaction. However, some are also detected in taste organs like proboscises and legs. These results suggest that some A-IRs in H. armigera likely bear a dual function with their involvement in olfaction and gustation. Results from mating experiments show that two HarmIRs (IR1.2 and IR75d) expression is significantly up-regulated in antennae of mated female moths. However, no expression difference is observed between unmated female and male adults, suggesting an association with female host-searching behaviors. Our current study has greatly extended the IR gene repertoire resource in Lepidoptera, and more importantly, identifies potential IR candidates for olfactory, gustatory and oviposition behaviors in the cotton bollworm.

Secure genome-wide association analysis using multiparty computation.

Most sequenced genomes are currently stored in strict access-controlled repositories1–3. Free access to these data could improve the power of genome-wide association studies (GWAS) to identify disease-causing genetic variants and may aid in the discovery of new drug targets4,5. However, concerns over genetic data privacy6–9 may deter individuals from contributing their genomes to scientific studies10 and in many cases, prevent researchers from sharing data with the scientific community11. Although several cryptographic techniques for secure data analysis exist12–14, none scales to computationally intensive analyses, such as GWAS. Here we describe an end-to-end protocol for large-scale genome-wide analysis that facilitates quality control and population stratification correction in 9K, 13K, and 23K individuals while maintaining the confidentiality of underlying genotypes and phenotypes. We show the protocol could feasibly scale to a million individuals. This approach may help to make currently restricted data available to the scientific community and could potentially enable ‘secure genome crowdsourcing,’ allowing individuals to contribute their genomes to a study without compromising their privacy.

Genetic study links components of the autonomous nervous system to heart-rate profile during exercise

Heart rate (HR) responds to exercise by increasing during exercise and recovering after exercise. As such, HR is an important predictor of mortality that researchers believe is modulated by the autonomic nervous system. However, the mechanistic basis underlying inter-individual differences has yet to be explained. Here, we perform a large-scale genome-wide analysis of HR increase and HR recovery in 58,818 UK Biobank individuals. Twenty-five independent SNPs in 23 loci are identified to be associated (p < 8.3 × 10−9) with HR increase or HR recovery. A total of 36 candidate causal genes are prioritized that are enriched for pathways related to neuron biology. No evidence is found of a causal relationship with mortality or cardiovascular diseases. However, a nominal association with parental lifespan requires further study. In conclusion, the findings provide new biological and clinical insight into the mechanistic underpinnings of HR response to exercise. The results also underscore the role of the autonomous nervous system in HR recovery.

Harnessing the genome for characterization of G-protein coupled receptors in cancer pathogenesis.

G-protein coupled receptors (GPCRs) mediate numerous physiological processes and represent the targets for a vast array of therapeutics for diseases ranging from depression to hypertension to reflux. Despite the recognition that GPCRs can act as oncogenes and tumour suppressors by regulating oncogenic signalling networks, few drugs targeting GPCRs are utilized in cancer therapy. Recent large-scale genome-wide analyses of multiple human tumours have uncovered novel GPCRs altered in cancer. However, work aiming to determine which GPCRs from these lists are the drivers of tumourigenesis, and hence valid therapeutic targets, comprises a formidable challenge. The present review highlights recent studies providing evidence that GPCRs are relevant targets for cancer therapy through their effects on known cancer signalling pathways, tumour progression, invasion and metastasis, and the microenvironment. Furthermore, the review also explores how genomic analysis is beginning to highlight GPCRs as therapeutic targets in the age of personalized medicine.

Genome-wide association analysis of red blood cell traits in African Americans: the COGENT Network.

Laboratory red blood cell (RBC) measurements are clinically important, heritable and differ among ethnic groups. To identify genetic variants that contribute to RBC phenotypes in African Americans (AAs), we conducted a genome-wide association study in up to ~16 500 AAs. The alpha-globin locus on chromosome 16pter [lead SNP rs13335629 in ITFG3 gene; P < 1E-13 for hemoglobin (Hgb), RBC count, mean corpuscular volume (MCV), MCH and MCHC] and the G6PD locus on Xq28 [lead SNP rs1050828; P < 1E - 13 for Hgb, hematocrit (Hct), MCV, RBC count and red cell distribution width (RDW)] were each associated with multiple RBC traits. At the alpha-globin region, both the common African 3.7 kb deletion and common single nucleotide polymorphisms (SNPs) appear to contribute independently to RBC phenotypes among AAs. In the 2p21 region, we identified a novel variant of PRKCE distinctly associated with Hct in AAs. In a genome-wide admixture mapping scan, local European ancestry at the 6p22 region containing HFE and LRRC16A was associated with higher Hgb. LRRC16A has been previously associated with the platelet count and mean platelet volume in AAs, but not with Hgb. Finally, we extended to AAs the findings of association of erythrocyte traits with several loci previously reported in Europeans and/or Asians, including CD164 and HBS1L-MYB. In summary, this large-scale genome-wide analysis in AAs has extended the importance of several RBC-associated genetic loci to AAs and identified allelic heterogeneity and pleiotropy at several previously known genetic loci associated with blood cell traits in AAs.

A Genome-Wide Analysis of RNA Pseudoknots That Stimulate Efficient −1 Ribosomal Frameshifting or Readthrough in Animal Viruses

Programmed −1 ribosomal frameshifting (PRF) and stop codon readthrough are two translational recoding mechanisms utilized by some RNA viruses to express their structural and enzymatic proteins at a defined ratio. Efficient recoding usually requires an RNA pseudoknot located several nucleotides downstream from the recoding site. To assess the strategic importance of the recoding pseudoknots, we have carried out a large scale genome-wide analysis in which we used an in-house developed program to detect all possible H-type pseudoknots within the genomic mRNAs of 81 animal viruses. Pseudoknots are detected downstream from ~85% of the recoding sites, including many previously unknown pseudoknots. ~78% of the recoding pseudoknots are the most stable pseudoknot within the viral genomes. However, they are not as strong as some designed pseudoknots that exhibit roadblocking effect on the translating ribosome. Strong roadblocking pseudoknots are not detected within the viral genomes. These results indicate that the decoding pseudoknots have evolved to possess optimal stability for efficient recoding. We also found that the sequence at the gag-pol frameshift junction of HIV1 harbors potential elaborated pseudoknots encompassing the frameshift site. A novel mechanism is proposed for possible involvement of the elaborated pseudoknots in the HIV1 PRF event.

Novel Autism Subtype-Dependent Genetic Variants Are Revealed by Quantitative Trait and Subphenotype Association Analyses of Published GWAS Data

The heterogeneity of symptoms associated with autism spectrum disorders (ASDs) has presented a significant challenge to genetic analyses. Even when associations with genetic variants have been identified, it has been difficult to associate them with a specific trait or characteristic of autism. Here, we report that quantitative trait analyses of ASD symptoms combined with case-control association analyses using distinct ASD subphenotypes identified on the basis of symptomatic profiles result in the identification of highly significant associations with 18 novel single nucleotide polymorphisms (SNPs). The symptom categories included deficits in language usage, non-verbal communication, social development, and play skills, as well as insistence on sameness or ritualistic behaviors. Ten of the trait-associated SNPs, or quantitative trait loci (QTL), were associated with more than one subtype, providing partial replication of the identified QTL. Notably, none of the novel SNPs is located within an exonic region, suggesting that these hereditary components of ASDs are more likely related to gene regulatory processes (or gene expression) than to structural or functional changes in gene products. Seven of the QTL reside within intergenic chromosomal regions associated with rare copy number variants that have been previously reported in autistic samples. Pathway analyses of the genes associated with the QTL identified in this study implicate neurological functions and disorders associated with autism pathophysiology. This study underscores the advantage of incorporating both quantitative traits as well as subphenotypes into large-scale genome-wide analyses of complex disorders.

Advances and perspectives from genetic research: development of biological markers in Alzheimer’s disease

Despite important recent advances, a full understanding of the (genetic) etiology of Alzheimer’s disease (AD) is still a long way off. Large collaborative efforts are ongoing, as well as the exploration of various sources of genetic variation. Evidence supports the view that Mendelian early-onset familial forms of AD are caused by rare and usually highly penetrant mutations in three genes (APP, PSEN1 and PSEN2). Considering sporadic late-onset AD (LOAD), the APOE ε4 allele is by far the best-established risk gene. Recently published large-scale genome-wide analyses point to additionally relevant genetically associated loci, particularly CLU, PICALM and CR1. These susceptibility loci support existing hypotheses about the amyloid, lipid, chaperone and chronic inflammatory mechanisms in AD pathogenesis, and are therefore likely to provide the basis for the development of hypothesis-driven novel biomarker candidates. Additional genes, listed online in AlzGene (e.g., GAB2 or SORL1) have repeatedly shown risk effects in LOAD, and may be true risk genes, but this is much less certain. New epigenetic research provided some evidence that DNA modifications maybe involved in LOAD (e.g., post-mortem studies described both hypo- and hyper-methylation in AD-related susceptibility genes). With respect to biomarkers, elderly nondemented APOE ε4 carriers demonstrated distinct cerebrospinal fluid biomarker signatures and alterations of brain glucose metabolism similar to those observed in AD. Future research should evaluate the usefulness of newly detected AD risk genes and epigenetic changes as potential biomarkers towards genetic profiling of AD or for correlation with endophenotypes and therapeutic outcome.

Analysis of ancient human genomes

High-capacity sequencing technologies have dramatically reduced both the cost and time required to generate complete human genome sequences. Besides expanding our knowledge about existing diversity, the nature of these technologies makes it possible to extend knowledge in yet another dimension: time. Recently, the complete genome sequence of a 4,000-year-old human from the Saqqaq culture of Greenland was determined to 20-fold coverage. These data make it possible to investigate the population affinities of this enigmatic culture and, by identifying several phenotypic traits of this individual, provide a limited glimpse into how these people may have looked. While undoubtedly a milestone in ancient DNA research, the cost to generate an ancient genome, even from such an exceptionally preserved specimen, remains out of reach for most. Nonetheless, recently developed DNA capture methods, already applied to Neanderthal and fossil human mitochondrial DNA, may soon make large-scale genome-wide analysis of ancient human diversity a reality, providing a fresh look at human population history.

Genetic bases of urinary albumin excretion and related traits in hypertension

Epidemiological as well as animal studies have recognized the potential role of genetic factors in the development of microalbuminuria and related traits (renal insufficiency, end-stage renal disease and nephroangiosclerosis) in hypertension. To unravel genetic variants of susceptibility, candidate gene, linkage and genome wide scan analysis has been used. In spite of the great efforts that have been made in the field, sound knowledge about the major genetic variants causing the susceptibility to develop renal damage in hypertension is scarce, since many associations were not replicated or only showed association in a certain subgroup of patients. Looking initially at genes of the most important physiological pathways of blood pressure regulation, linkage and genome wide scan have also detected genes of the lipid metabolism and protein components of the glomerular structures as potential candidate genes. Well designed large-scale genome-wide analysis and replication studies with large cohorts can help in the future to clarify the genetic bases of renal damage in hypertension. Combined strategies can contribute toward a better understanding of the genetic basis of urinary albumin excretion and renal damage in hypertension

DNA Methylation Profiling Predicts Clinical Outcomes and Reveals Unique Insights Into the Molecular Complexity of Acute Myeloid Leukemia.

Abstract 707▪▪This icon denotes an abstract that is clinically relevant.Epigenetic deregulation of genes through aberrant DNA methylation has been widely reported in cancer. We hypothesized that in AML this aberrant DNA methylation does not occur randomly, but rather occurs in specific and distinct patterns. Therefore, large-scale genome-wide analysis of the DNA methylome could help explain and define the complexity underlying leukemia biology and reveal the existence of epigenetically defined variants of AML. Using the HELP microarray assay, which measures DNA methylation at 50,000 CpG sites annotated to ∼14,000 promoters, we obtained DNA methylation profiles for 344 AML patients seen at Erasmus University Medical Center. Median follow-up based on survivors was 18.2 months (7-215); median age: 48 years (15-77). Unsupervised analysis (hierarchical clustering, correlation distance with Ward's clustering method) demonstrated that based on their methylation profiles AML patients distributed into 16 cohorts. 11 of these groups were also defined by the presence of specific molecular lesions: inv(16) [cluster 1], t(8;21) [cluster 3], t(15;17) [cluster 6], CEBPA-mutant [clusters 4 and 9], CEBPA-silenced [cluster 10] NPM1-mutant [clusters 12, 13, 14 and 16] and 11q23 abnormalities [cluster 11]. Enrichment for cases harboring a specific molecular lesion within a given cluster was determined using Fisher's exact test (p<0.01). Additionally, 5 new AML subtypes were defined based on epigenetic profiling alone and had no other clinical or molecular feature in common. Kaplan-Meier survival analysis revealed a significant difference in overall survival (OS) between these novel AML subtypes: 2-year OS±SE; 58.8%±8.4% and 45.2%±8.9% for clusters 5 and 7, respectively, vs. 23.6%±5.7%, 26.4%±9.2% and 33.3%±13.6%, for clusters 2, 8 and 15, respectively (log rank test, p=0.04). After adjustment for age, cytogenetic risk, NPM1 and FLT3-ITD status in a multivariate Cox proportional hazards regression model including all the clusters with ≥ 10 patients, 4 of these 5 novel clusters presented a statistically significant increased hazard ratio compared to the favorable risk inv(16) cluster. In contrast, the clinical outcomes of patients in cluster 5 were not significantly different from favorable risk patients with inv(16). In order to identify the genes affected by aberrant DNA methylation for each cluster, we performed a supervised analysis comparing each of the 16 clusters to normal CD34+ bone marrow progenitors (n=8) using ANOVA followed by Dunnet post hoc test, and selected genes with adjusted p values <0.05 and a methylation change >30%. The DNA methylation signatures of each cluster featured involvement of distinct gene networks and DNA regulatory elements, and displayed distinct degrees of hyper or hypomethylation with respect to normal CD34+ bone marrow cells. Of note, in spite of the variation in methylation across the 16 clusters, we identified a set of 45 genes that were almost universally aberrantly methylated (in >70% cases and present in at least 10/16 cluster signatures). This common epigenetic signature included the tumor suppressor PDZD2, the nuclear import proteins IPO8 and TNPO3, PIAS2, a regulator of MAP kinase signaling, CDK8, and CSDA, a regulator of CSF2. Gene expression profiling of the same patients indicated that at least 50% of these genes were also aberrantly silenced compared to normal CD34+ cells. Finally, we randomly divided the 344-patient cohort into a training group of 200 patients, a test group (n=95) and an independent validation group (n=49), and using the Supervised Principal Components algorithm identified a 15-gene methylation classifier that was predictive of OS (p<0.009) and event free survival (p<0.013). Furthermore, after adjustment for age, cytogenetic risk, NPM1, FLT3 and CEBPA status in a multivariate analysis, this classifier remained an independent risk factor for OS (Hazard ratio 1.29, 95% CI: 1.11-1.49; p<0.001). In summary, we have i) demonstrated that unique and distinct DNA methylation patterns characterize distinct forms of AML; ii) identified novel, epigenetically defined subgroups of AML with distinct clinical behavior; iii) revealed the presence of a consistently aberrantly methylated signature across AML subtypes, with confirmed silencing of the genes involved; and iv) report a 15-gene methylation classifier predictive of OS, and confirmed as an independent risk factor when adjusted for known AML covariates. Disclosures:No relevant conflicts of interest to declare.