Welcome to the CBRG
The Computational Biology Research Group (CBRG) provides computing support for
bioinformatics analysis at the University
of Oxford. We have expertise in many aspects of bioinformatics (sequence analysis, microarrays, proteomics and integration).
We especially encourage collaborations that require writing custom software, bioinformatics tools and databases.
An account with the CBRG has many benefits and gives automatic access to a large number of molecular biology computing packages and
to numerous biological databases.
We are based at the Sir William Dunn School of Pathology and at the
Weatherall Institute of Molecular Medicine. Full details can be found on
the contact details page.
Maher GJ, McGowan SJ, Giannoulatou E, Verrill C, Goriely A, Wilkie AO
Visualizing the origins of selfish de novo mutations in individual seminiferous tubules of human testes.
Proc Natl Acad Sci U S A (2016) 113: 2454-9
» View abstract
De novo point mutations arise predominantly in the male germline and increase in frequency with age, but it has not previously been possible to locate specific, identifiable mutations directly within the seminiferous tubules of human testes. Using microdissection of tubules exhibiting altered expression of the spermatogonial markers MAGEA4, FGFR3, and phospho-AKT, whole genome amplification, and DNA sequencing, we establish an in situ strategy for discovery and analysis of pathogenic de novo mutations. In 14 testes from men aged 39-90 y, we identified 11 distinct gain-of-function mutations in five genes (fibroblast growth factor receptors FGFR2 and FGFR3, tyrosine phosphatase PTPN11, and RAS oncogene homologs HRAS and KRAS) from 16 of 22 tubules analyzed; all mutations have known associations with severe diseases, ranging from congenital or perinatal lethal disorders to somatically acquired cancers. These results support proposed selfish selection of spermatogonial mutations affecting growth factor receptor-RAS signaling, highlight its prevalence in older men, and enable direct visualization of the microscopic anatomy of elongated mutant clones.
Davies JO, Telenius JM, McGowan SJ, Roberts NA, Taylor S, Higgs DR, Hughes JR
Multiplexed analysis of chromosome conformation at vastly improved sensitivity.
Nat Methods (2015) 13: 74-80
» View abstract
Methods for analyzing chromosome conformation in mammalian cells are either low resolution or low throughput and are technically challenging. In next-generation (NG) Capture-C, we have redesigned the Capture-C method to achieve unprecedented levels of sensitivity and reproducibility. NG Capture-C can be used to analyze many genetic loci and samples simultaneously. High-resolution data can be produced with as few as 100,000 cells, and single-nucleotide polymorphisms can be used to generate allele-specific tracks. The method is straightforward to perform and should greatly facilitate the investigation of many questions related to gene regulation as well as the functional dissection of traits examined in genome-wide association studies.
Ramcharan R, Aleksic T, Kamdoum WP, Gao S, Pfister SX, Tanner J, Bridges E, Asher R, Watson AJ, Margison GP, Woodcock M, Repapi E, Li JL, Middleton MR, Macaulay VM
IGF-1R inhibition induces schedule-dependent sensitization of human melanoma to temozolomide.
Oncotarget (2015) 6: 39877-90
» View abstract
Prior studies implicate type 1 IGF receptor (IGF-1R) in mediating chemo-resistance. Here, we investigated whether IGF-1R influences response to temozolomide (TMZ), which generates DNA adducts that are removed by O6-methylguanine-DNA methyltransferase (MGMT), or persist causing replication-associated double-strand breaks (DSBs). Initial assessment in 10 melanoma cell lines revealed that TMZ resistance correlated with MGMT expression (r = 0.79, p = 0.009), and in MGMT-proficient cell lines, with phospho-IGF-1R (r = 0.81, p = 0.038), suggesting that TMZ resistance associates with IGF-1R activation. Next, effects of IGF-1R inhibitors (IGF-1Ri) AZ3801 and linsitinib (OSI-906) were tested on TMZ-sensitivity, cell cycle progression and DSB induction. IGF-1Ri sensitized BRAF wild-type and mutant melanoma cells to TMZ in vitro, an effect that was independent of MGMT. Cells harboring wild-type p53 were more sensitive to IGF-1Ri, and showed schedule-dependent chemo-sensitization that was most effective when IGF-1Ri followed TMZ. This sequence sensitized to clinically-achievable TMZ concentrations and enhanced TMZ-induced apoptosis. Simultaneous or prior IGF-1Ri caused less effective chemo-sensitization, associated with increased G1 population and reduced accumulation of TMZ-induced DSBs. Clinically relevant sequential (TMZ â†’ IGF-1Ri) treatment was tested in mice bearing A375M (V600E BRAF, wild-type p53) melanoma xenografts, achieving peak plasma/tumor IGF-1Ri levels comparable to clinical Cmax, and inducing extensive intratumoral apoptosis. TMZ or IGF-1Ri caused minor inhibition of tumor growth (gradient reduction 13%, 25% respectively), while combination treatment caused supra-additive growth delay (72%) that was significantly different from control (p < 0.01), TMZ (p < 0.01) and IGF-1Ri (p < 0.05) groups. These data highlight the importance of scheduling when combining IGF-1Ri and other targeted agents with drugs that induce replication-associated DNA damage.
Twigg SR, Forecki J, Goos JA, Richardson IC, Hoogeboom AJ, van den Ouweland AM, Swagemakers SM, Lequin MH, Van Antwerp D, McGowan SJ, Westbury I, Miller KA, Wall SA, van der Spek PJ, Mathijssen IM, Pauws E, Merzdorf CS, Wilkie AO
Gain-of-Function Mutations in ZIC1 Are Associated with Coronal Craniosynostosis and Learning Disability.
Am J Hum Genet (2015) 97: 378-88
» View abstract
Human ZIC1 (zinc finger protein of cerebellum 1), one of five homologs of the Drosophila pair-rule gene odd-paired, encodes a transcription factor previously implicated in vertebrate brain development. Heterozygous deletions of ZIC1 and its nearby paralog ZIC4 on chromosome 3q25.1 are associated with Dandy-Walker malformation of the cerebellum, and loss of the orthologous Zic1 gene in the mouse causes cerebellar hypoplasia and vertebral defects. We describe individuals from five families with heterozygous mutations located in the final (third) exon of ZIC1 (encoding four nonsense and one missense change) who have a distinct phenotype in which severe craniosynostosis, specifically involving the coronal sutures, and variable learning disability are the most characteristic features. The location of the nonsense mutations predicts escape of mutant ZIC1 transcripts from nonsense-mediated decay, which was confirmed in a cell line from an affected individual. Both nonsense and missense mutations are associated with altered and/or enhanced expression of a target gene, engrailed-2, in a Xenopus embryo assay. Analysis of mouse embryos revealed a localized domain of Zic1 expression at embryonic days 11.5-12.5 in a region overlapping the supraorbital regulatory center, which patterns the coronal suture. We conclude that the human mutations uncover a previously unsuspected role for Zic1 in early cranial suture development, potentially by regulating engrailed 1, which was previously shown to be critical for positioning of the murine coronal suture. The diagnosis of a ZIC1 mutation has significant implications for prognosis and we recommend genetic testing when common causes of coronal synostosis have been excluded.
Ledaki I, McIntyre A, Wigfield S, Buffa F, McGowan S, Baban D, Li JL, Harris AL
Carbonic anhydrase IX induction defines a heterogeneous cancer cell response to hypoxia and mediates stem cell-like properties and sensitivity to HDAC inhibition.
Oncotarget (2015) 6: 19413-27
» View abstract
Carbonic anhydrase IX (CAIX) is strongly induced by hypoxia and its overexpression is associated with poor therapeutic outcome in cancer. Here, we report that hypoxia promotes tumour heterogeneity through the epigenetic regulation of CAIX. Based on hypoxic CAIX expression we identify and characterize two distinct populations of tumour cells, one that has inducible expression of CAIX and one that does not. The CAIX+ve population is enriched with cells expressing cancer stem cell markers and which have high self-renewal capacity. We show that differential CAIX expression is due to differences in chromatin structure. To further investigate the relationship between chromatin organization and hypoxic induction of CAIX expression we investigated the effect of JQ1 an inhibitor of BET bromodomain proteins and A366 a selective inhibitor of the H3K9 methyltransferase G9a/GLP. We identified that these drugs were able to modulate hypoxic CAIX expression induction. This further highlights the role of epigenetic modification in adaption to hypoxia and also in regulation of heterogeneity of cells within tumours. Interestingly, we identified that the two subpopulations show a differential sensitivity to HDAC inhibitors, NaBu or SAHA, with the CAIX positive showing greater sensitivity to treatment. We propose that drugs modulating chromatin regulation of expression may be used to reduce heterogeneity induced by hypoxia and could in combination have significant clinical consequences.
Clynes D, Jelinska C, Xella B, Ayyub H, Scott C, Mitson M, Taylor S, Higgs DR, Gibbons RJ
Suppression of the alternative lengthening of telomere pathway by the chromatin remodelling factor ATRX.
Nat Commun (2015) 6: 7538
» View abstract
Fifteen per cent of cancers maintain telomere length independently of telomerase by the homologous recombination (HR)-associated alternative lengthening of telomeres (ALT) pathway. A unifying feature of these tumours are mutations in ATRX. Here we show that expression of ectopic ATRX triggers a suppression of the pathway and telomere shortening. Importantly ATRX-mediated ALT suppression is dependent on the histone chaperone DAXX. Re-expression of ATRX is associated with a reduction in replication fork stalling, a known trigger for HR and loss of MRN from telomeres. A G-quadruplex stabilizer partially reverses the effect of ATRX, inferring ATRX may normally facilitate replication through these sequences that, if they persist, promote ALT. We propose that defective telomere chromatinization through loss of ATRX promotes the persistence of aberrant DNA secondary structures, which in turn present a barrier to DNA replication, leading to replication fork stalling, collapse, HR and subsequent recombination-mediated telomere synthesis in ALT cancers.
Belaya K, Rodríguez Cruz PM, Liu WW, Maxwell S, McGowan S, Farrugia ME, Petty R, Walls TJ, Sedghi M, Basiri K, Yue WW, Sarkozy A, Bertoli M, Pitt M, Kennett R, Schaefer A, Bushby K, Parton M, Lochmüller H, Palace J, Muntoni F, Beeson D
Mutations in GMPPB cause congenital myasthenic syndrome and bridge myasthenic disorders with dystroglycanopathies.
Brain (2015) 138: 2493-504
» View abstract
Congenital myasthenic syndromes are inherited disorders that arise from impaired signal transmission at the neuromuscular junction. Mutations in at least 20 genes are known to lead to the onset of these conditions. Four of these, ALG2, ALG14, DPAGT1 and GFPT1, are involved in glycosylation. Here we identify a fifth glycosylation gene, GMPPB, where mutations cause congenital myasthenic syndrome. First, we identified recessive mutations in seven cases from five kinships defined as congenital myasthenic syndrome using decrement of compound muscle action potentials on repetitive nerve stimulation on electromyography. The mutations were present through the length of the GMPPB, and segregation, in silico analysis, exon trapping, cell transfection followed by western blots and immunostaining were used to determine pathogenicity. GMPPB congenital myasthenic syndrome cases show clinical features characteristic of congenital myasthenic syndrome subtypes that are due to defective glycosylation, with variable weakness of proximal limb muscle groups while facial and eye muscles are largely spared. However, patients with GMPPB congenital myasthenic syndrome had more prominent myopathic features that were detectable on muscle biopsies, electromyography, muscle magnetic resonance imaging, and through elevated serum creatine kinase levels. Mutations in GMPPB have recently been reported to lead to the onset of muscular dystrophy dystroglycanopathy. Analysis of four additional GMPPB-associated muscular dystrophy dystroglycanopathy cases by electromyography found that a defective neuromuscular junction component is not always present. Thus, we find mutations in GMPPB can lead to a wide spectrum of clinical features where deficit in neuromuscular transmission is the major component in a subset of cases. Clinical recognition of GMPPB-associated congenital myasthenic syndrome may be complicated by the presence of myopathic features, but correct diagnosis is important because affected individuals can respond to appropriate treatments.
Taylor JC, Martin HC, Lise S, Broxholme J, Cazier JB, Rimmer A, Kanapin A, Lunter G, Fiddy S, Allan C, Aricescu AR, Attar M, Babbs C, Becq J, Beeson D, Bento C, Bignell P, Blair E, Buckle VJ, Bull K, Cais O, Cario H, Chapel H, Copley RR, Cornall R, Craft J, Dahan K, Davenport EE, Dendrou C, Devuyst O, Fenwick AL, Flint J, Fugger L, Gilbert RD, Goriely A, Green A, Greger IH, Grocock R, Gruszczyk AV, Hastings R, Hatton E, Higgs D, Hill A, Holmes C, Howard M, Hughes L, Humburg P, Johnson D, Karpe F, Kingsbury Z, Kini U, Knight JC, Krohn J, Lamble S, Langman C, Lonie L, Luck J, McCarthy D, McGowan SJ, McMullin MF, Miller KA, Murray L, Németh AH, Nesbit MA, Nutt D, Ormondroyd E, Oturai AB, Pagnamenta A, Patel SY, Percy M, Petousi N, Piazza P, Piret SE, Polanco-Echeverry G, Popitsch N, Powrie F, Pugh C, Quek L, Robbins PA, Robson K, Russo A, Sahgal N, van Schouwenburg PA, Schuh A, Silverman E, Simmons A, Sørensen PS, Sweeney E, Taylor J, Thakker RV, Tomlinson I, Trebes A, Twigg SR, Uhlig HH, Vyas P, Vyse T, Wall SA, Watkins H, Whyte MP, Witty L, Wright B, Yau C, Buck D, Humphray S, Ratcliffe PJ, Bell JI, Wilkie AO, Bentley D, Donnelly P, McVean G
Factors influencing success of clinical genome sequencing across a broad spectrum of disorders.
Nat Genet (2015) 47: 717-26
» View abstract
To assess factors influencing the success of whole-genome sequencing for mainstream clinical diagnosis, we sequenced 217 individuals from 156 independent cases or families across a broad spectrum of disorders in whom previous screening had identified no pathogenic variants. We quantified the number of candidate variants identified using different strategies for variant calling, filtering, annotation and prioritization. We found that jointly calling variants across samples, filtering against both local and external databases, deploying multiple annotation tools and using familial transmission above biological plausibility contributed to accuracy. Overall, we identified disease-causing variants in 21% of cases, with the proportion increasing to 34% (23/68) for mendelian disorders and 57% (8/14) in family trios. We also discovered 32 potentially clinically actionable variants in 18 genes unrelated to the referral disorder, although only 4 were ultimately considered reportable. Our results demonstrate the value of genome sequencing for routine clinical diagnosis but also highlight many outstanding challenges.
Zhang Y, Makvandi-Nejad S, Qin L, Zhao Y, Zhang T, Wang L, Repapi E, Taylor S, McMichael A, Li N, Dong T, Wu H
Interferon-induced transmembrane protein-3 rs12252-C is associated with rapid progression of acute HIV-1 infection in Chinese MSM cohort.
AIDS (2015) 29: 889-94
» View abstract
The interferon-inducible transmembrane protein-3 (IFITM3) is a protein that restricts multiple pathogenic viruses such as influenza virus. The single-nucleotide polymorphism rs12252-C, which is rare in Caucasian populations, but much more common in the Han Chinese population, has been found in much higher homozygous frequency in patients with severe acute influenza. Until now, there has been no study on the effect of this genetic variant on the clinical control of other viral infections.To investigate the impact of IFITM3-rs12252 genotypes on primary HIV-1 infection progression in an acute HIV-1-infected cohort in Beijing (PRIMO), China.We identified IFITM3-rs12252 genotypes of 178 acute HIV-1-infected patients and 196 HIV-negative candidates from the PRIMO cohort. HIV-1 viral load and CD4(+) T-cell counts were monitored at multiple time points during the first year of infection, and the association between IFITM3-rs12252 genotype and disease progression was evaluated.The current study shows that the IFITM3-rs12252 genetic variant affects the progression of HIV-1 infection, but not the acquisition. A significantly higher frequency of the CC/CT genotypes was found in rapid progressors compared to nonprogressors. Patients with CC/CT genotypes showed an elevated peak viremia level and significantly lower CD4(+) T-cell count at multiple time points during the first year of primary infection, and a significantly higher risk of rapid decline of the CD4(+) T-cell count to below 350â€Š cells/Î¼l.A novel association between IFITM3 gene polymorphism and rapid disease progression is reported in an acute HIV-1-infected MSM cohort in China.
French A, Yang CT, Taylor S, Watt SM, Carpenter L
Human induced pluripotent stem cell-derived B lymphocytes express sIgM and can be generated via a hemogenic endothelium intermediate.
Stem Cells Dev (2015) 24: 1082-95
» View abstract
The differentiation of human pluripotent stem cells to the B-cell lymphoid lineage has important clinical applications that include in vitro modeling of developmental lymphogenesis in health and disease. Here, we first demonstrate the capacity of human induced pluripotent stem cells (hiPSCs) to differentiate into CD144(+)CD73(-)CD43/CD235a(-) cells, characterized as hemogenic endothelium, and show that this population is capable of differentiating to CD10(+)CD19(+) B lymphocytes. We also demonstrate that B lymphocytes generated from hiPSCs are able to undergo full VDJ rearrangement and express surface IgM (sIgM(+)), thus representing an immature B-cell subset. Efficiency of sIgM expression on the hiPSC-derived B lymphocytes (âˆ¼ 5% of CD19(+) cells) was comparable with B lymphocytes generated from human umbilical cord blood (UCB) hematopoietic progenitor cells. Importantly, when assessed by global transcriptional profiling, hiPSC-derived B-cells show a very high level of similarity when compared with their UCB-derived counterparts, such that from more than 47,000 different transcripts, only 45 were significantly different (with a criteria adjusted P value P<0.05, log FC >1.5 or 2.8-fold). This represents a unique in vitro model to delineate critical events during lymphogeneisis in development and lymphoid diseases such as acute lymphocytic leukemia.
Babbs C, Lloyd D, Pagnamenta AT, Twigg SR, Green J, McGowan SJ, Mirza G, Naples R, Sharma VP, Volpi EV, Buckle VJ, Wall SA, Knight SJ, Parr JR, Wilkie AO
De novo and rare inherited mutations implicate the transcriptional coregulator TCF20/SPBP in autism spectrum disorder.
J Med Genet (2014) 51: 737-47
» View abstract
Autism spectrum disorders (ASDs) are common and have a strong genetic basis, yet the cause of âˆ¼70-80% ASDs remains unknown. By clinical cytogenetic testing, we identified a family in which two brothers had ASD, mild intellectual disability and a chromosome 22 pericentric inversion, not detected in either parent, indicating de novo mutation with parental germinal mosaicism. We hypothesised that the rearrangement was causative of their ASD and localised the chromosome 22 breakpoints.The rearrangement was characterised using fluorescence in situ hybridisation, Southern blotting, inverse PCR and dideoxy-sequencing. Open reading frames and intron/exon boundaries of the two physically disrupted genes identified, TCF20 and TNRC6B, were sequenced in 342 families (260 multiplex and 82 simplex) ascertained by the International Molecular Genetic Study of Autism Consortium (IMGSAC).IMGSAC family screening identified a de novo missense mutation of TCF20 in a single case and significant association of a different missense mutation of TCF20 with ASD in three further families. Through exome sequencing in another project, we independently identified a de novo frameshifting mutation of TCF20 in a woman with ASD and moderate intellectual disability. We did not identify a significant association of TNRC6B mutations with ASD.TCF20 encodes a transcriptional coregulator (also termed SPBP) that is structurally and functionally related to RAI1, the critical dosage-sensitive protein implicated in the behavioural phenotypes of the Smith-Magenis and Potocki-Lupski 17p11.2 deletion/duplication syndromes, in which ASD is frequently diagnosed. This study provides the first evidence that mutations in TCF20 are also associated with ASD.
Bassett AR, Azzam G, Wheatley L, Tibbit C, Rajakumar T, McGowan S, Stanger N, Ewels PA, Taylor S, Ponting CP, Liu JL, Sauka-Spengler T, Fulga TA
Understanding functional miRNA-target interactions in vivo by site-specific genome engineering.
Nat Commun (2014) 5: 4640
» View abstract
MicroRNA (miRNA) target recognition is largely dictated by short 'seed' sequences, and single miRNAs therefore have the potential to regulate a large number of genes. Understanding the contribution of specific miRNA-target interactions to the regulation of biological processes in vivo remains challenging. Here we use transcription activator-like effector nuclease (TALEN) and clustered regularly interspaced short palindromic repeat (CRISPR)/Cas9 technologies to interrogate the functional relevance of predicted miRNA response elements (MREs) to post-transcriptional silencing in zebrafish and Drosophila. We also demonstrate an effective strategy that uses CRISPR-mediated homology-directed repair with short oligonucleotide donors for the assessment of MRE activity in human cells. These methods facilitate analysis of the direct phenotypic consequences resulting from blocking specific miRNA-MRE interactions at any point during development.
Armitage AE, Stacey AR, Giannoulatou E, Marshall E, Sturges P, Chatha K, Smith NM, Huang X, Xu X, Pasricha SR, Li N, Wu H, Webster C, Prentice AM, Pellegrino P, Williams I, Norris PJ, Drakesmith H, Borrow P
Distinct patterns of hepcidin and iron regulation during HIV-1, HBV, and HCV infections.
Proc Natl Acad Sci U S A (2014) 111: 12187-92
» View abstract
During HIV type-1 (HIV-1), hepatitis C virus (HCV), and hepatitis B virus (HBV) infections, altered iron balance correlates with morbidity. The liver-produced hormone hepcidin dictates systemic iron homeostasis. We measured hepcidin, iron parameters, cytokines, and inflammatory markers in three cohorts: plasma donors who developed acute HIV-1, HBV, or HCV viremia during the course of donations; HIV-1-positive individuals progressing from early to chronic infection; and chronically HIV-1-infected individuals (receiving antiretroviral therapy or untreated). Hepcidin increased and plasma iron decreased during acute HIV-1 infection, as viremia was initially detected. In patients transitioning from early to chronic HIV-1 infection, hepcidin in the first 60 d of infection positively correlated with the later plasma viral load set-point. Hepcidin remained elevated in individuals with untreated chronic HIV-1 infection and in subjects on ART. In contrast to HIV-1, there was no evidence of hepcidin up-regulation or hypoferremia during the primary viremic phases of HCV or HBV infection; serum iron marginally increased during acute HBV infection. In conclusion, hepcidin induction is part of the pathogenically important systemic inflammatory cascade triggered during HIV-1 infection and may contribute to the establishment and maintenance of viral set-point, which is a strong predictor of progression to AIDS and death. However, distinct patterns of hepcidin and iron regulation occur during different viral infections that have particular tissue tropisms and elicit different systemic inflammatory responses. The hypoferremia of acute infection is therefore a pathogen-specific, not universal, phenomenon.
Taylor S, Noble R
HTML5 PivotViewer: high-throughput visualization and querying of image data on the web.
Bioinformatics (2014) 30: 2691-2
» View abstract
Visualization and analysis of large numbers of biological images has generated a bottle neck in research. We present HTML5 PivotViewer, a novel, open source, platform-independent viewer making use of the latest web technologies that allows seamless access to images and associated metadata for each image. This provides a powerful method to allow end users to mine their data.Documentation, examples and links to the software are available from http://www.cbrg.ox.ac.uk/data/pivotviewer/. The software is licensed under GPLv2.
Woll PS, Kjällquist U, Chowdhury O, Doolittle H, Wedge DC, Thongjuea S, Erlandsson R, Ngara M, Anderson K, Deng Q, Mead AJ, Stenson L, Giustacchini A, Duarte S, Giannoulatou E, Taylor S, Karimi M, Scharenberg C, Mortera-Blanco T, Macaulay IC, Clark SA, Dybedal I, Josefsen D, Fenaux P, Hokland P, Holm MS, Cazzola M, Malcovati L, Tauro S, Bowen D, Boultwood J, Pellagatti A, Pimanda JE, Unnikrishnan A, Vyas P, Göhring G, Schlegelberger B, Tobiasson M, Kvalheim G, Constantinescu SN, Nerlov C, Nilsson L, Campbell PJ, Sandberg R, Papaemmanuil E, Hellström-Lindberg E, Linnarsson S, Jacobsen SE
Myelodysplastic syndromes are propagated by rare and distinct human cancer stem cells in vivo.
Cancer Cell (2014) 25: 794-808
» View abstract
Evidence for distinct human cancer stem cells (CSCs) remains contentious and the degree to which different cancer cells contribute to propagating malignancies in patients remains unexplored. In low- to intermediate-risk myelodysplastic syndromes (MDS), we establish the existence of rare multipotent MDS stem cells (MDS-SCs), and their hierarchical relationship to lineage-restricted MDS progenitors. All identified somatically acquired genetic lesions were backtracked to distinct MDS-SCs, establishing their distinct MDS-propagating function in vivo. In isolated del(5q)-MDS, acquisition of del(5q) preceded diverse recurrent driver mutations. Sequential analysis in del(5q)-MDS revealed genetic evolution in MDS-SCs and MDS-progenitors prior to leukemic transformation. These findings provide definitive evidence for rare human MDS-SCs in vivo, with extensive implications for the targeting of the cells required and sufficient for MDS-propagation.