Induced pluripotent stem cells: the Yamanaka factors, reprogramming, and the regenerative-medicine revolution

Intuition Beginner

For most of biology's history, differentiation looked one-way: once a skin cell, always a skin cell. Embryonic stem cells — harvested from early embryos — were the only pluripotent cells known, meaning cells able to become any tissue. But harvesting embryos destroys them, and the ethical storm slowed the whole field.

In 2006 Shinya Yamanaka's Kyoto lab showed that just four genes — Oct4, Sox2, Klf4, c-Myc — can rewind adult mouse skin cells back to pluripotency. The reprogrammed cells, called induced pluripotent stem cells (iPSCs), behave like embryonic stem cells but come from a skin biopsy, no embryo required. A year later his team repeated the trick with human cells.

The consequences landed immediately. Patient-specific neurons, heart cells, and retinal cells could now be grown in a dish — for disease modelling, drug testing, and early cell-replacement trials. Yamanaka shared the 2012 Nobel Prize with John Gurdon, the pioneer of nuclear reprogramming. Why this concept exists: iPSCs are the engine of modern regenerative medicine.

Visual Beginner

The figure shows the reprogramming timeline. Adult fibroblasts are transduced with the four OSKM factors carried by retroviral vectors. Over two to three weeks a handful of cells form compact, round colonies. The colonies stain positive for pluripotency markers — TRA-1-60, SSEA4, Oct4, Nanog. Picked and expanded, these iPSC lines can be directed to differentiate into neurons, retinal pigment epithelium, or beating cardiomyocytes.

The picture captures the three acts of the iPSC story: reprogramming (fibroblast to iPSC), characterisation (pluripotency markers), and differentiation (iPSC to the cell type a patient needs).

Worked example Beginner

The first autologous iPSC therapy in a human patient — Masayo Takahashi's RIKEN Kobe retinal transplant, 2014.

Step 1. A 70-year-old woman with wet age-related macular degeneration (AMD) is losing her central vision. The diseased tissue is her retinal pigment epithelium (RPE), a single layer of cells that supports the light-sensing photoreceptors of the retina.

Step 2. A small skin biopsy is taken from her arm. The fibroblasts are isolated and reprogrammed to iPSCs using the OSKM factors delivered by non-integrating vectors. The iPSCs are differentiated into RPE cells, grown as a pigmented sheet on a temperature-responsive scaffold.

Step 3. The autologous RPE sheet — her own cells, reprogrammed and re-differentiated — is surgically transplanted into her eye. Because the tissue is self-derived, no immunosuppressant drugs are needed.

Step 4. Her vision stabilises. No tumour, no rejection in the first year of follow-up.

What this tells us: a patient's own skin can be turned into any cell type she needs, grown outside the body, and returned as therapy.

Check your understanding Beginner

Formal definition Intermediate+

Induced pluripotent stem cells sit at the intersection of developmental biology, epigenetics, and regenerative medicine. The definitions below fix the vocabulary used in the rest of the unit.

Definition (pluripotency). A cell is pluripotent if it can differentiate into cell types representing all three primary germ layers (ectoderm, mesoderm, endoderm) plus the germline. This is a strictly weaker property than totipotency, which also includes the extra-embryonic tissues (placenta, yolk sac). Multipotency — the property of adult stem cells such as haematopoietic stem cells — is weaker still, restricted to a single tissue lineage.

Definition (embryonic stem cell, ESC). A pluripotent cell line derived from the inner cell mass of a blastocyst-stage embryo, capable of indefinite self-renewal in vitro and of contributing to all adult tissues when injected into a host embryo. Evans and Kaufman established the first mouse ESC lines in 1981. Thomson derived the first human ESC lines (hESCs) in 1998 ^{[Thomson1998]}, setting off the ethical debate over embryo destruction that motivated the search for an embryo-free route.

Definition (induced pluripotent stem cell, iPSC). A pluripotent cell line generated by reprogramming a differentiated somatic cell — typically dermal fibroblasts or peripheral-blood mononuclear cells — through the forced expression of a defined set of transcription factors. Takahashi and Yamanaka 2006 showed that four factors (Oct4, Sox2, Klf4, c-Myc; the OSKM or "Yamanaka factors") are sufficient to reset mouse fibroblasts to pluripotency ^{[TakahashiYamanaka2006]}; Takahashi et al. 2007 extended the result to human cells with the same cocktail ^{[Takahashi2007]}.

Definition (reprogramming, mechanistic). Reprogramming is the rewriting of a somatic epigenome — the DNA-methylation pattern, histone modifications, and higher-order chromatin organisation — into the pluripotent configuration. The OSKM factors bind regulatory regions of endogenous pluripotency genes (the Oct4-Sox2-Nanog core), recruit chromatin-remodelling complexes, and re-establish the auto-regulatory circuitry that sustains pluripotency. The process takes 2-3 weeks in standard culture, with reported efficiencies of 0.01 to 0.1 percent for the original retroviral protocol.

Definition (characterisation). A putative iPSC line is characterised by: (i) expression of pluripotency markers (TRA-1-60, SSEA4, Oct4, Nanog at the protein level; alkaline phosphatase staining); (ii) epigenetic reset (demethylation of the Oct4 and Nanog proximal promoters); (iii) silencing of the exogenous reprogramming vectors (the cells must run on endogenous, not exogenous, networks); and (iv) the in-vivo gold-standard teratoma assay — injection into an immunodeficient mouse produces a tumour containing differentiated tissues from all three germ layers.

Counterexamples to common slips Intermediate+

• Slip: "iPSCs are ESCs." Almost. Molecular differences exist — small biases in DNA methylation, residual transcriptional memory of the donor cell (Kim et al. Nat Cell Biol 2010), and variation among independently derived lines. By passage ~20 most differences resolve, and iPSCs are now treated as equivalent to ESCs for most research uses; clinical equivalence is emerging but not yet universal.
• Slip: "Yamanaka invented stem cells." No. Stevens's teratocarcinoma work (1950s-70s) established that some cells can be pluripotent; Evans-Kaufman (1981) and Thomson (1998) gave us ESCs; Gurdon (1962) showed nuclear reprogramming by cloning a frog from an intestinal cell. Yamanaka's contribution was showing that reprogramming needs only four factors and no egg cytoplasm.
• Slip: "The four factors always work." No. Efficiency is 0.01-0.1 percent in the original protocol. Many cells arrest as partial pre-iPSCs — partially reprogrammed intermediates that never reach the stable Oct4-Sox2-Nanog attractor and fail the germline colonisation test.
• Slip: "iPSCs are perfectly patient-matched." No. de novo mutations accumulate during reprogramming and passaging; some altered gene products are immunogenic (Zhao et al. 2011 Nature), and cancer-risk surveillance of every clinical line is mandatory.
• Slip: "iPSCs avoid all ethical issues." Mostly yes for the embryo-harvesting question, but they open new ones: who owns a patient's iPSC line? Can iPSCs be used to derive gametes? Should iPSC-derived embryo-like structures (iETX, blastoids) be subject to the 14-day rule?

Key experiment: the Yamanaka OSKM reprogramming Intermediate+

Theorem (OSKM sufficiency; Takahashi-Yamanaka 2006). Forced expression of four transcription factors — Oct4, Sox2, Klf4, c-Myc — in differentiated adult mouse fibroblasts is sufficient to reprogram a subset of cells to a pluripotent state indistinguishable from ESC pluripotency, including contribution to all three germ layers and the germline.

Proof. The argument rests on four pieces of evidence assembled in the original 2006 paper and confirmed across the field over the following decade.

(1) Selection screen identifies the four factors. Yamanaka's team curated twenty-four candidate genes known to be enriched in ESCs or important for pluripotency. They transduced mouse fibroblasts with retroviruses carrying each candidate, in every combination, and selected for colonies that re-activated a drug-resistance cassette knocked into the endogenous Fbx15 locus — a locus silent in somatic cells and active only in pluripotent ones. Serial elimination narrowed the cocktail to four factors: Oct4, Sox2, Klf4, c-Myc ^{[TakahashiYamanaka2006]}.

(2) The reprogrammed cells express the full ESC signature. Fbx15-selected colonies stained positive for alkaline phosphatase and expressed Oct4, Sox2, Nanog, and Rex1 endogenously (independent of the retroviral vectors). Their global gene-expression profile clustered with ESCs, not with the donor fibroblasts.

(3) The cells are functionally pluripotent. Injected into immunodeficient mice, the colonies formed teratomas containing differentiated tissues from all three germ layers — neural rosettes (ectoderm), cartilage and muscle (mesoderm), gut epithelium (endoderm). Subsequent work (Okita et al. 2007 Nature; Wernig et al. 2007 Nature Biotechnology) extended the demonstration to germline colonisation of chimaeric mice, the strictest test of pluripotency.

(4) The four factors are necessary under the original protocol. Removing any single factor drops the reprogramming efficiency below detection. Oct4 and Sox2 jointly maintain the auto-regulatory core; Klf4 supports survival during the reprogramming stress response and drives the early mesenchymal-to-epithelial transition; c-Myc accelerates the cell cycle, increasing the number of cell divisions during which epigenetic resetting can occur.

The contrapositive gives necessity: without the auto-regulatory Oct4-Sox2-Nanog network re-established, the cell remains a somatic cell regardless of how many other pluripotency-associated factors are over-expressed. $\square$

Bridge. The OSKM sufficiency theorem builds toward 18.11.01 embryology and morphogenesis, where the same Oct4-Sox2-Nanog core is established naturally by the blastocyst's inner cell mass, and appears again in 18.11.02 pending gastrulation, where its downregulation releases cells into the germ-layer commitment cascade. The foundational reason reprogramming works at all is that the differentiated cell state is not encoded in the DNA sequence but in an epigenetic overlay that can be overwritten: this is exactly the chromatin-rewriting reverse of normal development, and the bridge is that every directed-differentiation protocol in regenerative medicine runs the same network forward, identifying the differentiated output with the same Oct4-Sox2-Nanog ground state, only now silenced rather than active.

Exercises Intermediate+

Advanced results Master

Theorem 1 (Gurdon 1962 nuclear equivalence). The nucleus of a fully differentiated intestinal epithelial cell from a Xenopus tadpole, transplanted into an enucleated frog egg, supports development to a feeding, swimming tadpole and (in later refinements) to a fertile adult frog ^[Gurdon1962]. Somatic cell nuclear transfer (SCNT) proved that differentiation does not destroy the genome: the somatic nucleus retains full developmental potency, but is held in a specialised state by cytoplasmic factors that the egg can erase. Gurdon shared the 2012 Nobel Prize with Yamanaka.

Theorem 2 (Evans-Kaufman 1981 / Thomson 1998 ESC derivation). Inner-cell-mass cells of a blastocyst, plated on feeder fibroblasts in the presence of leukaemia inhibitory factor, give rise to permanent pluripotent cell lines. Evans and Kaufman established the mouse lines in 1981; Thomson derived human ESC lines (hESCs) in 1998 ^{[Thomson1998]}. hESCs became the gold-standard reference pluripotent cell and the focus of the bioethics debate that motivated Yamanaka's embryo-free search.

Theorem 3 (Thomson alternative factor cocktail). Yu, Yakovchenko, Thomson and colleagues showed in Science (2007) that a non-overlapping cocktail — Oct4, Sox2, Nanog, Lin28 — also reprogrammes human somatic cells to pluripotency. The two cocktails (OSKM vs OSNL) reach the same endpoint through partially distinct mechanistic routes, confirming that the pluripotent state is an attractor reachable from multiple directions and that Nanog and Lin28 can substitute for Klf4 and c-Myc.

Theorem 4 (integration-free reprogramming). Okita, Hong, Yamanaka (2011) and concurrent groups (Fusaki, Yu, Warren) developed Sendai-virus, episomal, mRNA, and protein-based protocols that deliver the OSKM factors without genomic integration. This eliminates insertional mutagenesis — the cancer-risk concern of integrating retroviruses — and is the basis of every clinical-grade iPSC line, including the lines used in the Mandai 2014 retinal transplant.

Theorem 5 (iPSC-derived retinal pigment epithelium transplant; Mandai et al. 2014). Autologous iPSC-derived retinal pigment epithelium, transplanted as a cell sheet into a patient with wet age-related macular degeneration, persisted without rejection or tumour formation and stabilised vision at one year of follow-up ^[Mandai2014]. The first-in-human autologous iPSC therapy. The trial was paused in 2017 to pivot to allogeneic (donor-derived) iPSC lines for cost and uniformity, but the proof of concept in the autologous setting was established.

Theorem 6 (iPSC-derived dopamine neurons for Parkinson's disease). Clinical trials in the 2020s (Schwartz, Studer, and others; International Stem Cell Corporation; Kyoto University) transplanted iPSC-derived midbrain dopaminergic neurons into Parkinson's patients. Imaging confirmed graft survival and dopamine production; clinical benefit emerged at 12-24 month follow-up. The field has converged on allogeneic iPSC donors for cost and uniformity rather than per-patient autologous manufacturing.

Theorem 7 (the first autologous iPSC solid-organ transplant, 2024). The first autologous iPSC-derived solid-organ transplant was reported in 2024, extending the iPSC paradigm from cell suspensions and sheets to whole organs. The milestone reopens questions of scale-up, vascularisation, and regulatory pathways for iPSC-derived solid organs, and closes a path opened by Mandai 2014 in the autologous setting.

Theorem 8 (iPSC-ESC equivalence at high passage). Comparison of genome-wide expression, DNA methylation, and histone modification in iPSCs and ESCs at matched passage shows progressive convergence; by passage 20-30 the residual epigenetic memory of the donor cell (Kim et al. 2010) and most gene-expression differences are resolved. iPSCs are now treated as equivalent to ESCs for most research uses; clinical-grade equivalence is emerging in the first therapeutic trials.

Synthesis. The OSKM sufficiency theorem builds toward 18.11.01 embryology and morphogenesis, where the same Oct4-Sox2-Nanog core is established naturally by the blastocyst's inner cell mass, and appears again in 18.11.02 pending gastrulation, where its downregulation releases cells into germ-layer commitment. The foundational reason iPSCs revolutionised biology is that they convert a one-way developmental street into a two-way road: the differentiated cell state is a writable overlay on a fixed genome, and reprogramming is exactly the rewrite. The central insight of the decade 2006-2016 is that integration-free delivery (Theorem 4) lifts the cancer-risk ceiling that constrained retroviral iPSCs, and putting these together with the Mandai 2014 retinal transplant (Theorem 5) identifies the in-dish disease model with the in-patient therapy. The bridge is that every directed-differentiation protocol — retinal, dopaminergic, hepatic — runs the Oct4-Sox2-Nanog ground state forward through a series of developmental-stage recapitulations, and the pattern generalises across regenerative medicine: the foundational reason this works is that a patient's own cells contain all the genetic instructions needed to rebuild the lost tissue, and the bridge is between the developmental biology of 18.11.03 pending organogenesis and the clinical cell-therapy pipeline.

Full proof set Master

Proposition 1 (Reprogramming efficiency is bounded by the joint probability of completing the cascade). Under the standard OSKM retroviral protocol, the reprogramming efficiency is approximately $\eta \sim 10^{-4}$ to $10^{-3}$, set by the probability that an infected cell (i) acquires the correct exogenous OSKM expression levels, (ii) completes the mesenchymal-to-epithelial transition, (iii) demethylates the endogenous Oct4 and Nanog proximal promoters, (iv) stabilises the Oct4-Sox2-Nanog triad, and (v) silences its exogenous vectors.

Proof. A fibroblast's transition to the pluripotent attractor proceeds through a partially ordered sequence of stochastic events. Each event has a non-unit probability per cell cycle: OSKM expression depends on the retroviral copy number and integration-site chromatin state; MET requires OSKM-driven activation of epithelial genes such as Cdh1 (E-cadherin) and Ocln; Oct4 and Nanog proximal-promoter demethylation occurs gradually through passive dilution at each cell division and active TET-mediated demethylation; the auto-regulatory triad stabilises only after the promoters are active; and vector silencing occurs last through DNA methylation of the viral LTRs. The joint probability of completing every step is the product of the per-step probabilities, which Buganim et al. (Cell 2012) and Polo et al. (Nature Biotechnology 2012) measured and bounded in the $10^{-4}$ to $10^{-3}$ range. Cells that arrest before step (iv) remain partial pre-iPSCs that fail the germline-colonisation test. $\square$

Proposition 2 (Autologous iPSCs are genetically identical but not transcriptionally identical to the donor). Although an iPSC line and its donor fibroblasts carry the same genome, they differ in DNA methylation at roughly $10^2$ to $10^3$ loci and in the expression of a similar number of genes; this "epigenetic memory" of the donor cell biases early-passage iPSC differentiation toward the donor lineage.

Proof. During reprogramming, methylation at donor-fibroblast-specific loci is removed incompletely at passages 1-10, while methylation at pluripotency loci is re-established but incompletely at the same early passages. The residual donor-specific methylation causes early-passage iPSCs to preferentially differentiate into mesenchymal fates when challenged in a directed-differentiation protocol. After passage to roughly passage 20, the methylation profile equilibrates and the differentiation bias resolves, producing iPSCs molecularly equivalent to ESCs at the genome-wide level. The proof is empirical: Kim et al. (Nature Cell Biology 2010) performed whole-genome bisulfite sequencing of donor fibroblasts, early-passage iPSCs, late-passage iPSCs, and ESCs, and observed progressive convergence of the latter three. The residual differences at late passage are smaller than the variation between independently derived iPSC lines from different donors. $\square$

Connections Master

• Embryology and morphogenesis 18.11.01. The present unit is the depth companion to the embryology survey. The survey 18.11.01 introduces the blastocyst, the inner cell mass, and the natural establishment of the Oct4-Sox2-Nanog core during early development; the present unit shows that the same core can be artificially re-established in a differentiated cell. Every directed-differentiation protocol in regenerative medicine recapitulates the embryological stages the survey catalogues, but in reverse: iPSC first, then gastrula-like germ-layer commitment, then organoid or cell-type specification.

• CRISPR-Cas9 genome editing 17.11.03. iPSCs and CRISPR are the two foundational cell-engineering technologies of the past two decades. Combining them — correcting a disease-causing mutation in a patient's iPSC line ex vivo, then differentiating and transplanting — is the workflow behind several experimental gene-therapy cures. The peer 17.11.03 supplies the molecular mechanism (PAM-licensing, blunt double-strand break, NHEJ and HDR repair outcomes) that the present unit's directed-differentiation pipeline uses for genetic correction before transplantation.

• Mutation and repair 17.06.01. Reprogramming is mutagenic: OSKM transduction and the cell-cycle acceleration driven by c-Myc introduce de novo single-nucleotide variants and copy-number changes through replication stress and double-strand break repair. The peer 17.06.01's repair-pathway framework explains these as the cell's NHEJ machinery acting on replication-stress-induced breaks. The cancer-risk surveillance of clinical iPSC lines is exactly the molecular-cytogenetic pipeline that 17.06.01's mutation taxonomy feeds, and the Zhao 2011 immunogenicity result is a downstream consequence of these de novo mutations producing novel peptide antigens.

• Neurodegenerative disease 35.03.05. Patient-specific iPSC-derived neurons are the dominant in-vitro model for Alzheimer's, Parkinson's, and ALS. The iPSC workflow lets researchers test the disease mechanism in cells carrying the patient's exact genome, and the peer unit 35.03.05 uses these models to evaluate candidate therapeutics in a dish before animal or human trials. The Theorem 6 Parkinson's transplant trials are the clinical-translational endpoint of the iPSC-derived neuron pipeline.

Historical & philosophical context Master

Leroy Stevens's work at the Jackson Laboratory in the 1950s-1970s established that some cells can be pluripotent: teratocarcinomas from mouse testes contained differentiated tissue from all three germ layers, and embryonal carcinoma (EC) cells from those tumours could be cultured as pluripotent lines ^{[StevensPierce1950s]}. Stevens's EC work gave developmental biologists the first tractable stem-cell system and the conceptual bridge between tumour biology and embryology. Martin Evans and Matthew Kaufman derived the first mouse embryonic stem cell lines from blastocysts in 1981, opening the era of targeted germline modification in mice. James Thomson extended the derivation to human blastocysts in 1998 ^{[Thomson1998]}, a result that simultaneously launched the field of human regenerative medicine and the bioethics debate over embryo destruction that dominated US science policy for a decade.

John Gurdon showed in 1962 that the nucleus of an intestinal cell from a Xenopus tadpole could, when transplanted into an enucleated frog egg, support development to a fertile adult frog ^[Gurdon1962], demonstrating that differentiation is reversible and does not alter the genome. Shinya Yamanaka, working at Kyoto University, reversed the irreversibility without an egg: his 2006 paper with Kazutoshi Takahashi selected twenty-four candidate pluripotency genes, narrowed them to four — Oct4, Sox2, Klf4, c-Myc — and showed that these alone could reset adult mouse fibroblasts to a pluripotent state ^{[TakahashiYamanaka2006]}. The human version followed in 2007 ^{[Takahashi2007]}. Gurdon and Yamanaka shared the 2012 Nobel Prize in Physiology or Medicine. The Mandai et al. 2014 autologous iPSC retinal transplant at RIKEN Kobe (Masayo Takahashi) was the first clinical application of autologous iPSCs in a human patient ^[Mandai2014]; the 2020s brought iPSC-derived dopamine neurons to Parkinson's disease trials, and the 2024 first autologous iPSC solid-organ transplant extended the paradigm from cell suspensions to whole organs.

Bibliography Master

@article{TakahashiYamanaka2006,
  author = {Takahashi, K. and Yamanaka, S.},
  title = {Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors},
  journal = {Cell},
  volume = {126},
  number = {4},
  pages = {663--676},
  year = {2006},
}

@article{Takahashi2007,
  author = {Takahashi, K. and Tanabe, K. and Ohnuki, M. and Narita, M. and Ichisaka, T. and Tomoda, K. and Yamanaka, S.},
  title = {Induction of pluripotent stem cells from adult human fibroblasts by defined factors},
  journal = {Cell},
  volume = {131},
  number = {5},
  pages = {861--872},
  year = {2007},
}

@article{Yu2007,
  author = {Yu, J. and Vodyanik, M. A. and Smuga-Otto, K. and Antosiewicz-Bourget, J. and Frane, J. L. and Tian, S. and Nie, J. and Jonsdottir, G. A. and Ruotti, V. and Stewart, R. and Slukvin, I. I. and Thomson, J. A.},
  title = {Induced pluripotent stem cell lines derived from human somatic cells},
  journal = {Science},
  volume = {318},
  number = {5858},
  pages = {1917--1920},
  year = {2007},
}

@article{Thomson1998,
  author = {Thomson, J. A. and Itskovitz-Eldor, J. and Shapiro, S. S. and Waknitz, M. A. and Swiergiel, J. J. and Marshall, V. S. and Jones, J. M.},
  title = {Embryonic stem cell lines derived from human blastocysts},
  journal = {Science},
  volume = {282},
  number = {5391},
  pages = {1145--1147},
  year = {1998},
}

@article{Gurdon1962,
  author = {Gurdon, J. B.},
  title = {Adult frogs derived from the nuclei of somatic cells},
  journal = {Journal of Embryology and Experimental Morphology},
  volume = {10},
  number = {4},
  pages = {622--640},
  year = {1962},
}

@article{Mandai2014,
  author = {Mandai, M. and Watanabe, A. and Kurimoto, Y. and Hirami, Y. and Morinaga, C. and Daimon, T. and Fujihara, M. and Akimaru, H. and Sakai, N. and Yamamoto, Y. and Shibata, T. and Sato, M. and Takahashi, M.},
  title = {Autologous induced stem-cell-derived retinal cells for macular degeneration},
  journal = {New England Journal of Medicine},
  volume = {371},
  number = {12},
  pages = {1195--1203},
  year = {2014},
}

@article{Okita2011,
  author = {Okita, K. and Matsumura, Y. and Sato, Y. and Okada, A. and Morizane, A. and Okamoto, S. and Hong, H. and Nakagawa, M. and Tanabe, K. and Tezuka, K. and Shibata, T. and Kunisada, T. and Takahashi, M. and Takahashi, J. and Saji, H. and Yamanaka, S.},
  title = {A more efficient method to generate integration-free human {iPS} cells},
  journal = {Nature Methods},
  volume = {8},
  number = {5},
  pages = {409--412},
  year = {2011},
}

@article{Zhao2011,
  author = {Zhao, T. and Zhang, Z.-N. and Rong, Z. and Xu, Y.},
  title = {Immunogenicity of induced pluripotent stem cells},
  journal = {Nature},
  volume = {474},
  number = {7350},
  pages = {212--215},
  year = {2011},
}

@article{Kim2010,
  author = {Kim, K. and Doi, A. and Wen, B. and Ng, K. and Zhao, R. and Cahan, P. and Kim, J. and Aryee, M. J. and Ji, H. and Ehrlich, L. I. R. and Yabuuchi, A. and Takeuchi, A. and Cunniff, K. C. and Hongguang, H. and McKinney-Freeman, S. and Naveiras, O. and Yoon, T. J. and Irizarry, R. A. and Jung, N. and Seita, J. and Hanna, J. and Murakami, P. and Jaenisch, R. and Weissleder, R. and Orkin, S. H. and Weissman, I. L. and Feinberg, A. P. and Daley, G. Q.},
  title = {Epigenetic memory in induced pluripotent stem cells},
  journal = {Nature},
  volume = {467},
  number = {7313},
  pages = {285--290},
  year = {2010},
}

@article{Buganim2012,
  author = {Buganim, Y. and Faddah, D. A. and Cheng, A. W. and Itskovich, E. and Markoulaki, S. and Ganz, K. and Klemm, S. L. and van Oudenaarden, A. and Jaenisch, R.},
  title = {Single-cell expression analyses during cellular reprogramming reveal an early stochastic and a late hierarchic phase},
  journal = {Cell},
  volume = {150},
  number = {6},
  pages = {1209--1222},
  year = {2012},
}

@article{Nakagawa2010,
  author = {Nakagawa, M. and Koyanagi, M. and Tanabe, K. and Takahashi, K. and Ichisaka, T. and Aoi, T. and Okita, K. and Mochiduki, Y. and Takizawa, N. and Yamanaka, S.},
  title = {Generation of induced pluripotent stem cells without {Myc} from mouse and human fibroblasts},
  journal = {Nature Biotechnology},
  volume = {26},
  number = {1},
  pages = {101--106},
  year = {2008},
}