David M. Bodine - US grants

All of the cells in the peripheral blood are generated from a small population of hematopoietic stem cells (HSC) through a process of proliferation and differentiation known as hematopoiesis. The hematopoietic differentiation program includes well-defined stages at which the progeny of HSC become restricted to specific fates. The goal of this project is to define specific molecular signatures associated with specific stages of hematopoiesis. Project 1. DNA methylation is an essential epigenetic mark that is required for normal development and hematopoietic differentiation. We characterized genome-wide DNA methylation in primary mouse HSC, Common Myeloid progenitors (CMP), and erythroblasts (ERY). We found that DNA methylation was most abundant in HSC, with a 40% reduction in the number of peaks present in CMP, and a 67% reduction in the number of peaks present in ERY. Over 97% of the peaks present in CMP and ERY were also present in HSC, while only 3% of peaks arise de novo during differentiation. These results demonstrate that the methylation profile of erythroblasts is present in HSC and that a specific genome-wide decrease in DNA methylation occurs during erythropoiesis. (Hogart A, Lichtenberg J, Ajay SS, Anderson S, NIH Intramural Sequencing Center, Margulies EH, Bodine DM. Genome-wide DNA methylation profiles in hematopoietic stem and progenitor cells reveal overrepresentation of ETS transcription factor binding sites. Genome Res. 22 (8) 1407-18, 2012.) The magnitude of the Methyl-Seq data required the development of a novel software package we call SigSeeker. We used SigSeeker in a horizontal analysis of methylation in monocytic cells before and after treatment with 5-aza-cytidine, a DNA methylation inhibitor. (Lichtenberg J, Hogart A, Battle S. Bodine DM. Discovery of motif-based regulatory signatures in whole genome methylation experiments. Bioinformatics Open Source Conference (BOSC), 2012, Long Beach USA.) We found that SigSeeker was equally capable of analyzing ChIP-Seq data, demonstrating the applicability of the tool to many large-scale systems biology problems. (Dworkin AM, Lee M, Lichtenberg J, Patel SJ, Gildea D, Sakthianadeswaren A, NISC Comparative Sequencing Program, Foote S, Wolfsberg TG, Bodine DM, Crawford NPS. The metastasis suppressor RRP1B interacts with TRIM28 and HP1α at the c-MYC locus to induce heterochromatinization and silencing of c-MYC expression. PNAS submitted.) Our current plans are to extend these studies to generate methylation profiles for at least two additional cell types, the Megakaryocyte/Erythroid progenitor (MEP; the progenitor that gives rise to both the erythroid and megakaryocytic platelet lineages) and fully differentiated megakaryocytes. Through these studies we anticipate that we will identify methylation changes that are specific to the erythroid and megakaryocytic (platelet) lineages. Project 2. Diamond Blackfan anemia is an inherited bone marrow failure syndrome characterized by a severe deficiency of red blood cells, despite the fact that all other hematopoietic lineages differentiate normally and are present in normal numbers. Mutations in 9 different ribosomal protein genes have been identified in multiple unrelated families, most resulting in haploinsufficiency. However, 40% of DBA-associated mutations are missense mutations that cause a single amino acid change. These can be divided into two classes. The first type of amino acid substitution (Class I) interferes with the proper folding of the ribosomal protein, resulting in endosomal degradation and functional haploinsufficiency. The second type (Class II) produces proteins that appear to be folded normally as they are stable in the cytoplasm. We hypothesized that the Class II mutations act by a dominant negative mechanism. To test this hypothesis, we generated a transgenic mouse model that expresses an RPS19 mutation at codon 62 (RPS19R62W). Conditional expression of RPS19R62W in hematopoietic cells resulted in anemia with reduced numbers of erythroid progenitors. We concluded that RPS19R62W is a dominant negative DBA mutation. (Devlin EE, Dacosta L, Mohandas N, Elliott G, Bodine DM. A transgenic mouse model demonstrates a dominant negative effect of a point mutation in the RPS19 gene associated with Diamond-Blackfan anemia. Blood. 116 (15): 2826-35, 2010.) Attempts by others to generate mouse models of DBA by knocking out the mouse Rps19 gene have been unsuccessful. Preliminary data suggests that the output of the normal mouse Rps19 allele increases to compensate for the loss of the second allele. RPS19s only function appears to be in the assembly of the small ribosomal subunit. We hypothesized that mild deficiencies in the levels of ribosomal proteins that serve additional functions might be better candidates for a mouse model of DBA. Mutations in the RPL11 gene account for 15% of known DBA mutations. In addition to its role in the assembly of the large ribosomal subunit, RPL11 also interacts with MDM2, p53 and the rRNA processing pathways, which may account for the significantly higher frequency of craniofacial and physical abnormalities seen in patients with RPL11 mutations. We are currently generating an Rpl11 conditional knockout mouse to test this hypothesis and to provide an in vivo system to study the mechanism of erythroid failure and to test potentially useful drugs. Project 3. The goal of this project is to provide a molecular diagnosis to each DBA patient. We hypothesized that copy number variations that cause the deletion of a ribosomal protein gene could cause DBA in patients with normal ribosomal protein gene sequences. We used genome-wide SNP array profiling to evaluate regions of copy variation in DBA patients. In our initial sampling, we identified deletions at known DBA-related ribosomal protein gene loci in 20% of 50 DBA patients. These data suggest that ribosomal protein gene deletions are more common than previously suspected and are a significant cause of DBA. (Farrar JE, Vlachos A, Atsidaftos E, Carlson-Donohoe H, Markello TC, Arceci RJ, Ellis SR, Lipton JM, Bodine DM. Ribosomal protein gene deletions in Diamond-Blackfan Anemia. Blood. 118 (26): 6943-51, 2011). Using a sensitive new method to detect mosaic copy number variations (Markello TC, Carlson-Donohoe H, Sincan M, Adams D, Bodine DM, Farrar JE, Vlachos A, Lipton JM, Auerbach AD, Ostrander EA, Chandrasekharappa SC, Boerkoel CF, Gahl WA. Sensitive quantification of mosaicism using high density SNP arrays and the cumulative distribution function. Mol Genet Metab. 105 (4): 665-71, 2012.), we also identified novel variable mosaic loss involving chromosomal regions containing known DBA gene regions in 5 patients from 4 different kindreds. To identify novel DBA mutations, we plan to expand our sequence analysis of those patients who have not had mutations in known DBA genes identified by sequencing and do not have copy number variations identified by SNP-CGH. We will enroll 4 unrelated family groups consisting of parents, a DBA proband and either an affected or unaffected sibling for exome capture sequencing. The inclusion of the parents and sibling will add power to the analysis of the exome sequence. After the first round of sequencing we will screen the remaining pool of undiagnosed DBA patients for any newly identified mutation(s) before enrolling a second set of 4 families who do not carry a known mutation.

Cooleys anemia (b-Thalassemia major) and Sickle Cell Disease are caused by mutations in the b-globin genes. Severe cases can be cured by Bone Marrow Transplantation, but with a significant risk of death due to graft rejection. Gene Therapy, in which a viral vector containing a b-globin gene is inserted into patient bone marrow cells, carries none of the risks of graft rejection and could be an alternative cure for Cooleys anemia or SCD. However, current vectors contain the powerful enhancer elements of the b-globin locus control region (LCR), which raises concerns that the integration of the virus near oncogenes could lead to leukemia. Our long-term goal is to design gene transfer vectors for the treatment of Cooleys anemia (b-Thalassemia major) and or Sickle Cell Disease that address the three challenges confronting gene therapy for this disease: therapeutic levels of expression that do not depend on enhancer elements in the viral vector or the transcription unit, resistance to position dependent gene silencing and transduction of 20% or more of human HSC. We hypothesize that self-inactivating (SIN) vector backbones containing enhancer independent globin transcription units that are pseudotyped with an envelope that recognizes abundant receptors on HSC would be the optimal system for gene therapy of these diseases. We have developed several new approaches to deal with these problems. Firstly, we have identified sequences that increase the level of expression from the ankyrin promoter 8-9 fold. Since the ankyrin promoter carries it's own barrier element, this promoter is resistant to gene silencing. Using the wild type sequence we found that the ankyrin promoter gave about 1/2 the level of mRNA needed to cure SCD and treat Cooley's anemia, so we are preparing to test the modified ankyrin promoter in animal models. We have also identified all of the regulatory elements in the Slc4a1 locus, a gene that is expressed at therapeutic levels in red blood cells. These include enhancer, enhancer blockers and barrier elements. Our Slc4a1hg-globin vectors expressed g-globin at levels of as high as 17% of the level of endogenous mouse a-globin expression. For our patient studies we will pseudotype our lentivirus vector with the Feline Leukemia Virus type C (FeLV-C) envelope that recognize an abundant receptor expressed on human HSC. The objective of this proposal is to test these novel vectors in mouse models and in cells from Cooleys anemia patients to determine whether this strategy safely and efficiently deliver and express therapeutic levels of g-globin in erythroid cells. We plan to pursue the following two specific aims: 1. Evaluate the efficiency of transduction and gene expression of lentiviral vectors containing either the ankyrin g-globin or an Slc4a1-driven human g-globin gene flanked with distinct barrier elements. The working hypothesis for this aim is that these lentiviruses will safely allow for high HSC transduction efficiency and drive therapeutic levels of hg-globin without the need for enhancers. 2. To test the best vector in primitive human stem and progenitor cells from Cooleys anemia or SCD patients using the fetal sheep transplantation model. We believe that these studies will translate directly into safe and effective gene therapy for the two most common inherited hemoglobinopathies. Specific Aim 1: Evaluate the efficiency of transduction and gene expression of lentiviral vectors containing an Slc4a1-driven human -globin gene flanked with distinct barrier elements. We have identified sequence modifications in the ankyrin promoter activity that increase expression and verified that these also increase expression in transgenic mice. The new promoter has bee added to our prototype vectors for evaluation in the mouse models of beta thalassemia and SCD. The levels of human g-globin mRNA and protein will be determined by RNase protection and HPLC respectively. For the Slc4a1 vectors our basic lentiviral design consists of a 1.7 kb mouse Slc4a1 promoter linked to the human g-globin (Slc4a1g-globin) gene flanked by distinct barrier elements from the Slc4a1 locus. The barrier-flanked Slc4a1-globin has been inserted into a HIV-1-based SIN lenitvirus, in which the 3LTR promoter and enhancer elements are deleted to preclude LTR-driven oncogene transcription. We are adding the Slc4a1 enhancer blockers to the flanking barriers to prevent the enhancers in the promoter from activating downstream genes Specific Aim 2: To test the best vector in primitive human stem and progenitor cells from Cooleys anemia or SCD patients using the fetal sheep transplantation model. To investigate the therapeutic potential of the Slc4a1-globin vectors in human -thalassemia, we will transduce human CD34 stem and progenitor cells obtained from bone marrow of b-thalassemia patients. We have previously shown that oncoretroviruses pseudotyped with the FeLV-C envelope transduce human sheep repopulating cells at much higher frequencies than the GaLV or VSV-G envelopes. We have adapted this envelope to package lentivirus vectors by deleting the R peptide in the FeLV-C envelope, similar to the successful strategies employed to adapt the MLV amphotropic envelope for packaging lentiviruses. Patient CD34 cells will be transduced and injected into preimmune fetal sheep between 55 and 60 days of gestation. The transplanted lambs will be brought to term, and two weeks after birth blood samples will be analyzed for the presence of human HSC-derived cells using an anti-human CD45 antibody. DNA extracted from human lymphoid and myeloid cells will be analyzed for integration of the provirus and integration sites. Human erythrocytes (CD45, Glycophorin A) will be isolated from peripheral blood by fluorescence activated cell sorting and the proportions of beta-like globin chains to alpha-like globin chains will be analyzed by HPLC. As the lambs age, bone marrow samples will be obtained for the generation of myeloid and erythroid colony forming cells.

There are three related projects in this program: Project 1: Whole exome sequencing to identify DBA mutations. Project 2: In vitro culture and differentiation of DBA patient cells. Project 3: Small molecule screening and mechanistic studies of DBA. Project 1. The goal of this project is to provide a molecular diagnosis to each DBA patient. The first part of our genotyping screen is an analysis of ribosomal RNA (rRNA) processing in cultured peripheral blood lymphocytes. We have developed very sensitive metrics that can identify defects in both the large and small subunit ribosomal proteins (Farrar JE, Quarello P, Fisher R, O'Brien KA, Aspesi A, Parrella S, Henson AL, Seidel NE, Atsidaftos E, Prakash S, Bari S, Garelli E, Arceci RJ, Dianzani I, Ramenghi U, Vlachos A, Lipton JM, Bodine DM, Ellis SR. Exploiting pre-rRNA processing in Diamond Blackfan anemia gene discovery and diagnosis. Am J Hematol. 89(10):985-91, 2014). This can inform and confirm the targeted sequencing performed on all patients with a preliminary diagnosis of DBA. If no mutations are detected by targeted sequencing we proceed to SNParray analysis of deletions. We previously showed that deletions of ribosomal protein genes was responsible for about 10% of all DBA mutations (Farrar JE, Vlachos A, Atsidaftos E, Carlson-Donohoe H, Markello TC, Arceci RJ, Ellis SR, Lipton JM, Bodine DM. Ribosomal protein gene deletions in Diamond-Blackfan Anemia. Blood. 118 (26): 6943-51, 2011). These can include very small deletions that are well below the detection limit for cytogenetics (Vlachos A, Farrar JE, Atsidaftos E, Muir E, Narla A, Markello TC, Singh SA, Landowski M, Gazda HT, Blanc L, Liu JM, Ellis SR, Arceci RJ, Ebert BL, Bodine DM, Lipton JM. Diminutive somatic deletions in the 5q region lead to a phenotype atypical of classical 5q- syndrome. Blood. 122(14):2487-90, 2013). If no deletions are detected, we perform whole exome sequencing analysis of family groups consisting of parents, a DBA proband and either an affected or unaffected sibling. The inclusion of the parents and sibling added a great deal of power to our analysis allowing us to identify 2 strong candidate genes (MCM-2 and SEMA7A) in our first 8 families. We are currently validating these mutations in vitro and we are screening a pool of 96 undiagnosed DBA patients to see if they have any of the newly identified mutations. We have enrolled additional sets of families whose sequence is proceeding through the pipeline. However, our analysis of whole exome sequencing in the other 6 families we have completed has not identified an obvious candidate gene. We then hypothesized that the defect may lie outside the coding regions of the genome, and may involve deletions too small to be detected by SNP array. For families who have negative results with exome sequencing we have initiated 3 whole genome analyses designed to identify non-coding mutations, with an emphasis on deletions. Project 2: The goal of this project is to understand the specific lesions in DBA patients that prevent the generation of erythroid cells. We designed a two-step, 14 day in vitro culture system to generate erythroid cells from CD34+ stem/progenitor cells isolated from <10ml of peripheral blood (PB) collected from 12 patients enrolled in the DBAR. Patient cells were compared to healthy control PB CD34+ cells. At day 14, we routinely obtain at least 10x fewer CD235+ erythroid cells (proerythroblasts and basophilic erythroblasts) in DBA cultures vs. controls (1x106 vs. 1x107 from 1x104 CD34+ cells). Further, a 2-7 day delay in the acquisition of the erythroid cell surface marker CD235 was observed. Gene expression analyses was performed by Affymetrix GeneChip Human Gene ST Arrays and RNASeq to analyze protein coding and long non-coding RNA transcripts in cells from day 14 of erythroid cultures. Ingenuity pathway analysis (IPA) of the dysregulated genes between patients with ribosomal mutations, GATA1 mutations and controls revealed that many of the genes dysregulated in DBA were involved in multiple leukocyte migration and inflammatory signaling pathways, including the IL8, IL1R1, CXCR4, ICAM3, MPO, TNFSF10, and TLR4 genes with IL6, TNF, and lipopolysaccharide as top upstream regulators. Notably, the dysregulated genes in GATA1 patient cells largely overlapped that of the DBA patients with ribosomal mutations, including disruption of the leukocyte migration and inflammatory response genes. Patients with GATA1 and ribosomal protein mutations shared a number of dysregulated erythroid genes including AHSP, FAM132B, HEMGN, and TRIM10, however GATA1 patient cells showed GATA1 as the top upstream regulator and additionally showed dysregulation of heme biosynthesis pathway genes, including the ALAS2, FECH, CPOX, PPOX, and UROS. We are currently investigating the inflammatory pathways in DBA to identify novel targets for therapeutic development. Project 3. The goal of this project is to identify small molecules that can be used to treat DBA. The majority of DBA patients are heterozygous for a ribosomal protein gene carrying a deleterious mutation and a normal healthy gene. Haploinsufficeincy results because of decreased translation of functional RP protein. The RP mRNAs are among those RNAs with Terminal Oligo Pyrimidine (TOP) sequences in the 5UTR that regulates the rate of translation. We have developed a reporter cell line that has a luciferase gene with a TOP UTR. This cell line shows deceased translation of luciferase when treated with rapamycin, a known inhibitor of translation. In addition, serum starvation specifically inhibits the translation of TOP mRNAs. Our reporter cell line shows a 3-5-fold reduction in luciferase levels when serum starved. In collaboration with NCATS we are currently optimizing our reporter assay to fit a 1536 well plate format. Our plan is to treat serum starved cells with compounds and assay for increased luciferase output. We will screen small molecules to see which ones can increase the translation of TOP mRNA. These compounds will be further tested in patient cells and our mouse model (Devlin EE, Dacosta L, Mohandas N, Elliott G, Bodine DM. A transgenic mouse model demonstrates a dominant negative effect of a point mutation in the RPS19 gene associated with Diamond-Blackfan anemia. Blood. 116(15):2826-35, 2010.

Summary of Core projects active during FY2013: 1. Characterization of gene function using WISH and morpholinos: This year, we analyzed the function of the following candidate genes identified by genomic approaches using WISH and morpholinos: cecr1b, gne, itga11, sccpdha, vps45 and zak. In all cases, gene-specific morpholinos were designed to target translation start site and/or splicing junctions. Microinjections were performed to determine the optimal dose of each morpholino, followed by microinjections for phenotype analysis using imaging, WISH and transgenic lines. In some cases, activity of the wildtype and mutant human mRNAs was assesses using complementation of the morpholino phenotype. 2. Generation of knockout mutants using ZFNs and TALENs: In collaboration with the NHGRI Genomics Core, we developed high throughput method of fluorescent PCR for founder screening and genotyping to complete these projects in a cost-effective and timely fashion. We have generated multiple mutant alleles for ten genes and additional six genes are in progress. A brief summary of the various projects using targeted mutagenesis is given below: 2.1 Role of AK2 in reticular dysgenesis: To model reticular dysgenesis in zebrafish and investigate the mechanisms underlying its pathophysiology, we generated multiple ak2 mutants: a missense mutation (L124P) by TILLING and null mutants by ZFNs. Dr. Rissone has performed a comprehensive study of the effects of ak2 deficiency on hematopoiesis using morpholinos and both mutants, demonstrating a critical role of ak2 in erythroid development during primitive hematopoiesis and hematopoietic stem cell (HSC) development during definitive hematopoiesis. 2.2 Role of c-Met signaling in the posterior lateral line development: To investigate the role of met in the migrating lateral line primordium, we generated loss-of-function alleles using ZFNs that are being characterized by the Burgess lab. 2.3 Genes involved in regulation of hematopoiesis: To assist Liu lab in their investigation of the regulatory network involved in hematopoiesis, we have generated loss-of-function alleles using ZFNs and TALENs for sox17, runx1 and cbfb. Liu lab is currently analyzing expression of specific hematopoietic markers by WISH and transgenic lines in these mutants. 2.4 Modeling metabolic diseases in zebrafish: To model disorders of vitamin B12 metabolism, Cobalamin C (cblC) disease and methylmalonic academia, for insights into their underlying pathology and assessment of the efficacy of potential therapeutics, we generated loss-of-function alleles in mmachc and mut using ZFNs. Their characterization is currently underway by Venditti lab. 2.5 Candidate genes identified by genomic approaches: We have generated knockout mutants for the following genes identified as putative candidate genes in various studies undergoing at NHGRI: kctd7 (progressive myoclonic epilepsy), hint3 (distal myopathy with weakness and atrophy of distal muscles and high arched palate), ubqln4 (neurological phenotype), aifm2 (new immune disorder), pus3 (Kabuki-like syndrome), and bmp3 (craniofacial anomalies). Mutant generation for additional candidate genes is in progress. 3. Generation of stable transgenic lines: We have generated stable transgenic lines using tol2 mediated transgenesis to analyze effects of over-expression and specific mutations in hint3 and bmp3 genes. In addition, we are generating transgenic lines for use in cell tracking experiments by marking nuclei and membranes by photo-convertible fluorophores. 4. Evaluation of conserved non-coding sequences for enhancer activity using tol2 mediated transgenesis and ZED vector: Evaluation of highly conserved non-coding sequence variants, often detected as putative mutations and associations in human genetic diseases is challenging. Furthermore, computationally identified conserved elements are presumed to function as enhancers. Zebrafish provides an ideal in vivo system to test their role as enhancers by tol2 transgenesis and a minimum promoter plus the fragment to be tested driving GFP. ZED vector with RFP expression in muscle at 48hpf as an internal control of transgenesis was developed recently (Bessa et al., Developmental Dynamics 238: 2409-2417, 2009). We are currently evaluating wildtype and variant forms of two conserved elements associated with idiopathic scoliosis and sagittal craniosynostosis. Transient expression is variable due to mosaicism, therefore, we generated stable lines for the conserved element localized in the IRX cluster region and identified founders with germline transmission. 5. Chemical screening: The Core is facilitating characterization of the role of elg1 and msh6 in DNA repair by Myung laboratory using previously generated knockout fish lines and testing their sensitivity to various DNA damaging agents. In addition, we provide aliquots of the spectrum collection from Microsource Discovery Systems for chemical screening projects. 6. Education and outreach: Zebrafish embryos at different stages of development and transgenic lines with GFP showing circulating blood, blood vessels, or other specific organs make an attractive visual teaching tool. We have held many tours organized by the ITO, office of education, and NIH visitor center.

DESCRIPTION (provided by applicant): Defining the regulatory architecture of hematopoietic cells to elucidate lineage determination and differentiation can produce insights into developmental biology and can help identify targets with potential application to human diseases such as leukemias and anemias. Mouse hematopoiesis is a versatile system for studying gene regulation during differentiation because we can purify populations of progenitor and differentiated cells for genome-wide mapping of transcripts and regulatory sequences, and we can genetically manipulate critical proteins and cis-regulatory modules (CRMs) to study mechanisms of regulation. This application is for a renewal of a long-standing, productive collaboration among multiple investigators with complementary expertise in hematopoietic cell differentiation, gene regulation, genomics, bioinformatics and statistics. Our previous work laid a foundation of genome-wide data sets for transcriptomes, transcription factor occupancy and chromatin states in a cultured cell model for erythroid differentiation and in maturing primary cells in the erythroid and megakaryocytic lineages, which led to key new insights about regulation. We now propose to (Aim 1) generate genome-wide data on transcriptomes and informative epigenetic features in purified cells from each stage of differentiation from mouse hematopoietic stem cells to mature cells of the erythroid and myeloid lineages. For all cell types, including multilineage progenitor cells available only in small numbers, we propose to determine transcriptomes, DNA methylation, and chromatin accessibility (using a new method based on in vitro transposition). In more abundant cell types, we will use ChIP-seq to map transcription factors and histone modifications and also the chromosome conformation capture method Hi-C to build an interaction map of distal regulatory regions with target genes. We will then (Aim 2) conduct integrative, quantitative modeling to find genes differentially expressed and with different transcription factor binding patterns in the distinct lineages; within this set are candidates for genes involved in choice of cell lineage. A hypothesis-driven Bayesian network model will learn quantitative relationships between features, including expression level, and make predictions about how the system would behave after perturbation of both transcription factors and CRMs. We will then (Aim 3) conduct genetic manipulations to test hypotheses arising from integrative analysis in Aim 2. Specific hypotheses about genes involved in lineage choice will be tested by transduction of interfering or forced expression constructs into mouse fetal liver progenitor cells and bipotential cells in culture. Hypotheses from the quantitative modeling of determinants of levels of expression will be tested, targeting specific proteins (using transfections of cells with or withou GATA1) and CRMs (by Cas9-CRISPR-guided genome editing). The result of this proposed work will be deep, widely disseminated data on the regulatory landscape in multiple hematopoietic lineages and keener insights into how changes in regulatory proteins and chromatin lead to lineage choice and progressive differentiation.

Project Summary VISION: ValIdated Systematic IntegratiON of hematopoietic epigenomes Technological advances enabling the production of large numbers of rich, genome-wide, sequence-based datasets have transformed biology. However, the volume of data is overwhelming for most investigators. Also, we do not know the mechanisms by which the vast majority of epigenetic features regulate normal differentiation or lead to aberrant function in disease. We have formed an interdisciplinary, collaborative team of investigators to address the problem of how to effectively utilize the enormous amount of epigenetic data both for basic research and precision medicine. At this point, acquisition of data is no longer the major barrier to understanding mechanisms of gene regulation during normal and pathological tissue development. The chief challenges are how to: (i) integrate epigenetic data in terms that are accessible and understandable to a broad community of researchers, (ii) build validated quantitative models explaining how the dynamics of gene expression relates to epigenetic features, and (iii) translate information effectively from mouse models to potential applications in human health. These needs are addressed by the proposed ValIdated Systematic IntegratiON (VISION) of epigenetic data to analyze mouse and human hematopoiesis, a tractable system with clear clinical significance and importance to NIDDK. By pursuing the following Specific Aims, the interdisciplinary collaboration will deliver comprehensive catalogs of cis regulatory modules (CRMs), extensive chromatin interaction maps and deduced regulatory domains, validated quantitative models for gene regulation, and a guide for investigators to translate insights from mouse models to human clinical studies. These deliverables will be provided to the community in readily accessible, web-based platforms including customized genome browsers, databases with facile query interfaces, and data-driven on-line tools. Specifically, the proposed work in Aim 1 will build comprehensive, integrative catalogs of hematopoietic CRMs and transcriptomes by compiling and determining informative epigenetic features and transcript levels in hematopoietic stem and progenitor cells and in mature cells. CRMs will be predicted using the novel IDEAS (Integrative and Discriminative Epigenome Annotation System) method. Work proposed in Aim 2 will build and validate quantitative models for gene regulation informed by chromatin interaction maps and epigenetic data. Compiling and determining chromosome interaction frequencies will predict likely target genes for CRMs. Gene regulatory models will be built that predict the contributions of CRMs and specific proteins to regulated expression; these models will be validated by extensive testing using genome-editing in ten reference loci. Finally, work in Aim 3 will produce a guide for investigators to translate insights from mouse models to human clinical studies. This effort will include categorizing orthologous mouse and human genes by conservation versus divergence of expression patterns, assigning CRMs to informative categories of epigenomic evolution, and testing the interspecies functional maps experimentally by genome-editing.