- Research article
- Open Access
MPSS profiling of human embryonic stem cells
- Ralph Brandenberger†1,
- Irina Khrebtukova†3,
- R Scott Thies2,
- Takumi Miura1,
- Cai Jingli,
- Raj Puri4,
- Tom Vasicek3,
- Jane Lebkowski2 and
- Mahendra Rao1, 5Email author
© Brandenberger et al; licensee BioMed Central Ltd. 2004
- Received: 30 March 2004
- Accepted: 10 August 2004
- Published: 10 August 2004
Pooled human embryonic stem cells (hESC) cell lines were profiled to obtain a comprehensive list of genes common to undifferentiated human embryonic stem cells.
Pooled hESC lines were profiled to obtain a comprehensive list of genes common to human ES cells. Massively parallel signature sequencing (MPSS) of approximately three million signature tags (signatures) identified close to eleven thousand unique transcripts, of which approximately 25% were uncharacterised or novel genes. Expression of previously identified ES cell markers was confirmed and multiple genes not known to be expressed by ES cells were identified by comparing with public SAGE databases, EST libraries and parallel analysis by microarray and RT-PCR. Chromosomal mapping of expressed genes failed to identify major hotspots and confirmed expression of genes that map to the X and Y chromosome. Comparison with published data sets confirmed the validity of the analysis and the depth and power of MPSS.
Overall, our analysis provides a molecular signature of genes expressed by undifferentiated ES cells that can be used to monitor the state of ES cells isolated by different laboratories using independent methods and maintained under differing culture conditions
- Massively Parallel Signature Sequencing
- hESC Line
- UniGene Cluster
- Undifferentiated hESC
- Sage Library
Multiple large-scale analytical techniques to assess gene expression in defined cell populations have been developed. These include microarray analysis, EST enumeration, SAGE and MPSS. Each of these techniques offers unique advantages and disadvantages. Technique selection largely depends on the expertise of the investigator, the cost, the availability of the techniques, the amount of RNA/DNA that is available, and the existence of the genome databases. The human genome dataset is the best annotated one available [1, 2]- making large scale gene expression analysis of human tissues and cells uniquely fruitful for investigators due to the increased ability to identify full length transcripts with predicted gene function instead of EST's.
Human ES cells have been isolated relatively recently and ES cell genes are underrepresented in current databases. More importantly, recent evidence has suggested that mouse ES and human ES cells differ significantly in their fundamental biology [3, 4] and one cannot readily extrapolate from one species to another. However, comparing results between species may provide unique insights. Given the wealth of SAGE and microarray data available from rodent ES cells examining human ES cells with similar techniques as has been done recently by several investigators [3–11] should be very useful in furthering our understanding of this special stem cell population. Until recently however, it has been difficult to obtain RNA from a homogenous population of undifferentiated hESC for such an analysis as cells could not be grown without feeders and few unambiguous ES cell markers had been described. However, we and others have now described markers that will clearly assess the state of ES cells using a combination of immunocytochemistry and RT-PCR [3, 12, 13] In addition, techniques of harvesting ES cells away from feeder layers have been developed and verified (our unpublished results) and methods of growing ES cells without feeders have been described . These techniques, have allowed us (and others) to obtain large amounts of validated RNA/cDNA samples for comparison by microarray [3–11], SAGE  or EST enumeration .
We selected MPSS for this analysis as it offers some unique advantages over other methods including SAGE [15, 16]. MPSS offers sufficient depth of coverage when over one million transcripts are sequenced  and is efficient, as the numbers of sequences obtained are an order of magnitude larger than with shotgun sequencing or SAGE. It is relatively rapid with a turnaround of a six to ten weeks, and if done with human tissues, more than 80% of transcripts can be mapped to the human genome with current tools. Further, independent analysis has suggested that expression at greater than 3 tpm (transcripts per million) is predictive of detectable, reliable expression, equivalent to roughly one transcript per cell – a sensitivity that is unparalleled when compared to other large-scale analysis techniques . Finally, MPSS libraries can be translated into SAGE libraries and compared to existing SAGE library sets using freely available tools such as digital differential display, allowing ready comparisons to existing SAGE/MPSS libraries of mouse ES cells. It is important to note that we found 14 base pair SAGE tags are generally not as specific as 17 base MPSS signatures and that SAGE sampling depth is usually insufficient. Newer technologies such as extended sequencing to 20 base pairs in MPSS, 24 base pairs in SAGE or cheaper bead alternatives such as those described by Illumina may offer additional depth of coverage and a cheaper price but these at present remain limited in availability.
We have utilized MPSS using a pooled sample of three human ES cell lines grown in feeder-free culture conditions over multiple passages [17, 18] to assess the overall state of undifferentiated ES cells. Our rationale for using pooled sample rather than individual samples was based on the fact that no standardized medium and culture conditions have been established for growing and propagating ES cell lines. Variation observed by sampling single lines may be due to culture conditions rather than intrinsic differences. We reasoned therefore that a need existed to establish a reference baseline using pooled samples to enhance the similarities and provide evidence for candidate genes that should be examined for differences such as expression of HLA genes, Y chromosome and X chromosome genes, imprinted genes and genes regulating the methylation state. Our results show that MPSS provides a greater depth of coverage than EST scan or microarray and provides a comprehensive expression profile for this stem cell type. The data set generated allows us and others to identify multiple genes that were not previously known to be expressed in this population, including novel gene as well as obtain a global overview of pathways that are active during the process of self-renewal.
MPSSS analysis of pooled samples
Classification of the MPSS cDNA signatures. The signature classification used for annotation is shown * The Class 0 signatures are the signatures that hit genome more than 100 times, which is treated as a "repeat sequence". ** The polyA tail is defined as a stretch of A's (at least 13 out of 15 bases) that is no more than 50 bases away from the end of the source sequence. The polyA signal is either AATAAA or ATTAAA that has at least one base within the last 50 base before the end of the source sequence or the polyA tail. *** All the virtual signatures extracted from the genomic sequences are classified as class 1000 signatures.
Virtual Signature Class
Poly-Adenelation Features **
Either – Repeat Warning
Poly-A Signal, Poly-A Tail
Not 3' most
Not 3' most
Poly-A Signal, Poly-A Tail
Not 5' most
Not 5' most
Last before signal
Last before tail
Last in sequence
Not 3' most
Unknown – Derived from Genomic Sequence
Major pathways present at detectable levels by MPSS
Likewise, MPSS detected expression of several MHC Class I and II genes, suggesting that MPSS can identify differences between ES cell samples when HLA gene expression is used to type cells [17, 18]. We also note that both H19 and Igf2 were expressed at detectable levels. H19 and Igf2 are located adjacent to each other on chromosome 11p15.5 and are reciprocally regulated by imprinting, H19 being paternally imprinted, and IGF2 being maternally imprinted [23, 24]. It is therefore likely that their ratio of expression is likely to differ between cell populations and may represent a simple assessment of the imprinting status of cells.
Senesence and Aging related genes A subset of genes related to senescence and aging that may regulate the lack of senescense in ES cells is shown. Note that the telomerase, morf's, nortalins and sirtuins are all expressed in ES cells. *The TERT gene has a signature uniquely mapping to an intron (cryptic exon?), which was present in all runs of the ES cell analysis and was not found in other human samples (not shown).
IGF-/PTEN/Akt and Ras/Raf/MAP pathway A subset of genes related to Igf/PTEN pathway that are expressed in undifferentiated ES cells is shown. Note that the overall pattern of expression suggest that this pathway is active in undifferentiated ES cells.
ERK (p42 MAPK)
Novel genes enriched in hESC as assessed by MPSS A short list of genes of unknown function that are highly enriched in three ES cell lines comparing to 36 different tissues and cells are shown. A complete list of unknown genes expressed in pooled hESC cells is presented in supplementary tables. * NS-neural stem cells, TH-thymus, HY-hypothalamus, PG-pituitary gland, TE-testis ** this gene (Hs.507833 in the unigene Hs.169) is transcribed in antisense to HDCMA18P (Hs.278635)
Other 36, TPM*
CD250365:Homo sapiens transcribed sequence **
NM_019079:hypothetical protein FLJ10884
NM_018189:hypothetical protein FLJ10713
TH-47, HY-3, PG-3
AI636928:Homo sapiens transcribed sequences
CD174249:Homo sapiens transcribed sequence
AK092578:Sapiens cDNA FLJ35259 fis
BF223023:Homo sapiens transcribed sequences
NM_032805:hypothetical protein FLJ14549 (ZNF206)
XM_067369:similar to Heterochronic gene LIN-41
CD176172:Homo sapiens transcribed sequence
CD173198:Homo sapiens transcribed sequences
BU608353:Homo sapiens transcribed sequence
Comparing with other data sets
MPSS tpm of genes reported as enriched by microarray in hESC Table 5 Tpm of genes identified as overexpressed microarray analysis of six pooled human ES cell lines. Note that most of them have high tpm values and are detected by MPSS. * – PSIP2 and PSIP1 have 3' alternate termination and distinguished by MPSS (but not by microarray); ** – PODXL: TPM for signature of class 5; 3' most signature has double palindrome and underrepresented. *** – higher expression of GDF3 was detected in other ES cells (suppl.table for BG02 and not shown). **** – expression detected in other human samples (not shown).
Ribo 60S L30
Comparison with an EST scan analysis of 37,081 EST sequenced from a similar pooled sample of hESC [9, 10] also showed a high degree of concordance. The EST scan analysis detected 8,801 distinct UniGene clusters in hESC versus 9,996 distinct UniGene clusters expressed at 4 tpm or higher in the MPSS dataset. Of the 8,801 UniGene clusters identified by the EST scan, 1,139 are singletons, i.e. identified by only one EST out of the 37,081 total EST's. 5,286 UniGene clusters have 5 or more ESTs as evidence, and only 118 UniGene clusters have more than 100 EST's as evidence. In contrast, all 9,996 UniGene clusters identified by MPSS were detected at 4 or more tpm and identified in multiple sequencing runs. More than 8,000 have at least 10 tpm, and over 1,000 have more than 100 tpm. Thus, although the EST's are longer in length and thus easier to assign to a particular gene, MPSS appears more sensitive than EST scan. MPSS for example identified almost twice as many genes as EST scan consistent with the difference in the depth of analysis (No of sequences MPSS/EST).
Richards et al  have used SAGE analysis to two ES cell lines. Their analysis revealed expression of approximately four thousand genes which was significantly fewer than that identified by MPSS consistent with the fewer number of gene tags sequenced. Comparison of the data sets however showed good concordance particularly for genes expressed at higher tpm levels. The entire comparison is presented in supplementary table (additional file 18) and is available for download from Lynx . Overall MPSS could identify genes that other methods identified with an average concordance rate of 70%. The depth of analysis with MPSS at 2.4 million signatures however was significantly greater. MPSS in general identified many more genes than microarray or EST scan or SAGE (see above). The most direct comparison is with EST scan or SAGE, which do not rely on comparative gene expression to establish significance of gene expression. Overall our comparison suggests that MPSS results provide a complementary global overview of the transcriptome of the ES cell. The data supplement and extend the microarray, SAGE and EST scan data sets and provide an independent verification of the same. MPSS in addition identifies additional genes expressed particularly at lower tpm, that are either not present on microarrays or not detected with a lower resolution analysis.
Our results provide a global overview of the gene expression pattern of undifferentiated human ES cells and allow comparisons with other data sets. These results suggest the hESC are an actively dividing population of cells that exhibit high metabolic activity. Our analysis detected expression of approximately 10,600 unique transcripts, a figure that about a third of the total number of mapped genes. Unlike other cell types, however, a much larger fraction of unknown or novel genes was present. This high ratio likely represents the paucity of information available in existing libraries on this relatively newly characterized cell population rather than the possibility that ES cells use radically different pathways for self-renewal, survival, proliferation and differentiation.
Our results confirm the reported differences between rodent and human ES cells. We confirm the absence of expression of ERAS, Ehox and the orthologs PEPP1 and 2. The apparent lack of LIF requirement of hESC is reflected by the absence or low tpm levels for genes of the LIF pathway and high tpm for suppressors of LIF mediated signalling (see supporting information). The high level of expression of genes in the FGF pathway likely reflects the requirement of hESC for bFGF. The high level of FGFR1 expression suggests that FGFR1 is an important signal transducer and that FGF's other than FGF4 are important in hESC self-renewal. The high tpm of the fibronectin receptor also suggest that fibronectin or vitronectin are likely useful substitutes for matrigel and that activation of ras mediated signalling is likely critical, as has been described in the rodent ES cell analysis .
Comparing data from the MPSS analysis with microarray, SAGE and EST scan analyses suggest that MPSS is a powerful alternative to these techniques. MPSS identified virtually all of the genes highlighted as genes common between six different human ES cell lines surveyed by microarray. We noted that most genes detected by microarray were expressed at high tpm indicating that MPSS is more sensitive than microarray analysis. MPSS however appeared to be able to identify genes detected by microarray. Analysing an additional 400 markers detected by MPSS using focused microarray or RT-PCR confirmed their expression , (data not shown). Likewise, MPSS analysis showed good concordance with the EST scan data at a fraction of the price. In contrast to the EST scan, tpm levels determined by MPSS are highly correlated to the mRNA levels present in the cells, even at low tpm values , and (Lynx unpublished results). Due to the low sampling number of most EST scans, this is not true for relatively low number of EST's found for a particular gene, and can be used only as a rough estimate of gene expression. Unlike other in depth analyses, the absence of markers in MPSS runs is also a powerful control provided that the marker possesses a GATC site. The chromosomal distribution of the genes expressed in hESC did not reveal any bias for a particular chromosome or chromosomal region. While a couple of "hotspots" and several "cold spots" were identified, in no case was any region comprised of all transcribed or all silent genes.
Another important conclusion from our analysis is that selection of input RNA is critical. In our case we tested samples repeatedly to assess their purity and made considerable efforts to establish subclones that did not require feeder cells that could be potentially contribute transcripts to the analysis. Given the range of tpm of biologically relevant molecules (5 to 32,000 in this experiment) we predict that even a 5% contamination can confound results or detailed comparisons across different laboratories.
We note also that gene transcription from both the X and Y chromosome is observed indicating that at least subtle differences will exist between male and female lines even in the undifferentiated state. Sex-based gene expression, along with MHC gene expression and ratio of expression of imprinted genes could serve to distinguish between different ES cell populations. The present results further suggest that analysing embryoid bodies that differentiate stochastically or analysing tissue samples (with variable proportions of cells) by MPSS will prove more difficult and that results will be variable. We suggest that variability can be reduced by pooling samples, normalizing by careful testing for known markers of differentiation, by semi quantitative PCR, or by focused microarray analysis.
While MPSS is cost-effective and sensitive, it is by no means perfect. MPSS is limited by the requirement that DpnII sites (GATC) be present in a gene and be present in a unique locus such that the signature obtained is unique. For example, SNRF expression could not be assessed directly, as no GATC site is present. The signatures for ZFP42 are ambiguous and map to multiple transcripts. Although MPSS can distinguish between alternate transcript termination sites, MPSS cannot distinguish between alternative splicing events and possible incomplete digestion during the sample preparation process. Signature lengths are relatively short and it is possible to have to select between multiple genome hits (reviewed in . Sequencing is performed four bases at a time and transcripts that contain palindromic sequences (in particular double palindromes) are often undetected because of self-hybridization of single DNA strands on the bead. A survey of the genome suggests that this is a rare event (approximately 3% of all virtual signatures in human MGC database have double palindromes). The NODAL gene is an example for such an event, where the class 1 signature was lost and NODAL expression is detected only by a signature resulting from incomplete digestion during library construction (see results). The success of MPSS analyses also depends to a large extent on the quality of genomic information available and, in our opinion, currently is best utilized to analyse human cells. Furthermore, MPSS itself may not be the best method for routine, lower throughput analyses, given price per sample, sample processing time and the large amount of data generated, which requires considerable analysis. However, the database, once developed, is extremely valuable provided it is freely available to make comparisons and to select subsets of genes for further analysis. MPSS information can be effectively utilized by establishing a common database of markers expressed at a defined stage in the differentiation of cells. Additional data sets from sampling of cells at well-controlled stages of differentiation that can be readily accessed and compared to existing datasets will provide the most information while still being cost effective. The genome database is an example of such sharing that has proven to be an invaluable resource for our experiments. Such a strategy requires cooperative pooling of information and free sharing such that individual results can be readily compared against validated datasets. Our future experiments will be directed and developing additional data sets of ES cell differentiation, which can be shared in a manner similar to the present set.
Our results provide a comprehensive data set that can be effectively utilized to analyse expression patterns of known and unknown genes. Comparison with other data sets provides independent confirmation of results and shows a high level of concordance. The caveats to all such large-scale comparisons are discussed and the importance of pooling data and comparing across multiple data sets is demonstrated.
MPSS was performed using RNA from three pooled ES cell lines (H1, H7, and H9) that had been maintained in feeder free culture conditions and evaluated for the presence of ES cell markers and absence of markers of differentiation. The mRNA was converted to cDNA and digested with DpnII. The last DpnII site and the downstream 16 bases were cloned into Megaclone vectors and their sequences determined according to the MPSS protocol [15, 16, 25]. A total of 2.786.765 sequences were read from four different runs and 48,388 unique signatures were identified. The abundance for each signature was converted to transcripts per million (tpm) for the purpose of comparison between samples. Signatures at an abundance of less than 4 tpm or those that were not detected in at least two runs were removed and a total of 22,136 sequences were analyzed further. All data is available for download from Lynx 
MPSS signature classification and annotation
To generate a complete, annotated human signature database, we extracted all the possible signatures ("virtual signatures") from the human genome sequence, the human Unigene sequences, and human mitochondrion. Each virtual signature was ranked, as outlined in the table 1a, based on its position and orientation in the original sequence. Unigene, genomic, and mitochondrial hits were combined and grouped by signature. The annotation was then assigned to the signature in following order of preference: repeat warnings (signature hits more than 100 genome locations); mitochondrial hits; Unigene hits; genome hits (if no transcript match found). If a signature matched only one Unigene cluster, the MPSS signature class is the lowest class of the member sequences of the cluster. If a signature hits multiple Unigene clusters, the best cluster hit is selected based on the lowest MPSS signature class or the largest number of member sequences. The resulting signature database was used to annotate the data from the experiments Initially the signatures were annotated using genome version hg15 (April 2003, Golden Path, UCSC,) and Unigene build #161 (additional file 2). Recently we re-annotated all signatures using genome version hg16 (July 2003, Golden Path, UCSC) and Unigene build #169 (additional file 3). Both annotations are available for download in supplemental tables .
Analysis was performed as described in Bhattacharya et al.,  using six different samples. These included two lines from Bresagen (01 and 02), the pooled sample from Geron comprising feeder free subclones of (H1, H7, H9), H1, grown in our laboratory on feeders and H9 and I6 from Dr. Itskovitz-Eldor grown following their published protocols.
Chromosomal mapping of MPSS signatures and UniGene clusters to the human genome
MPSS signatures with a hit to a UniGene cluster were mapped to the Giemsa staining cytobands of the hg15 release of the human genome (April 10, 2003 freeze, ). By this method, 7731 MPSS signatures were mapped to the cytobands of the human genome. Similar mapping was done for all UniGene clusters for which the chromosomal mapping is known. In order to achieve a gene-based rather than a transcript (i.e. splice variant) based distribution of genes splice variants the UniGene clusters were filtered using LocusLink data , since LocusLink captures all characterized splice variants of a particular gene. 23,828 UniGene clusters were identified by this method and mapped to the cytobands of the human genome. To discover differences in the number of genes mapped to each cytoband, the number of genes mapped to each cytoband was compared to the total number of genes analyzed, for both the MPSS signatures as well as for the UniGene clusters. The Fisher test  was used to determine the statistical significance, using a p-value = 0.05 as cutoff.
Gene detection by RT-PCR
Total RNA was isolated from cell pellets using RNAeasy Qiagen mini protocol and kit. cDNA was synthesized using 100 ng of total RNA in a 20-μl reaction. Superscript II (Gibco-BRL), a modified Maloney murine leukemia virus RT, and Oligo (dT)12–18 primers were used according to the manufacturer's instructions (Gibco-BRL). The list of primers used for RT-PCR and annealing conditions are described previously ).
This work was supported by the NIA and an ALS center grant. JC, TM, MR were supported by the NIA. RP was supported by the FDA. ST, RB, JL are employees of Geron Inc. IK and TV are employees of Lynx Inc. We thank Drs. Ginis and Limke for careful manuscript reading and all members of our laboratories for constant stimulating discussions. MSR acknowledges the contributions of Dr. S. Rao that made undertaking this project possible.
- Gadiraju S, Vyhlidal CA, Leeder JS, Rogan PK: Genome-wide prediction, display and refinement of binding sites with information theory-based models. BMC Bioinformatics. 2003, 4: 38-10.1186/1471-2105-4-38.PubMed CentralView ArticlePubMedGoogle Scholar
- Camon E, Magrane M, Barrell D, Lee V, Dimmer E, Maslen J, Binns D, Harte N, Lopez R, Apweiler R, et al: The Gene Ontology Annotation (GOA) Database: sharing knowledge in Uniprot with Gene Ontology Genome-wide prediction, display and refinement of binding sites with information theory-based models. Nucleic Acids Res. 2004, 4: D262-6. 10.1093/nar/gkh021.View ArticleGoogle Scholar
- Ginis I, Luo Y, Miura T, Thies S, Brandenberger R, Gerecht-Nir S, Amit M, Hoke A, Carpenter MK, Itskovitz-Eldor J, et al: Differences between human and mouse embryonic stem cells. Dev Biol. 2004, 269: 360-80. 10.1016/j.ydbio.2003.12.034.View ArticlePubMedGoogle Scholar
- Pennacchio LA: Insights from human/mouse genome comparisons. Mamm Genome. 2003, 14: 429-36. 10.1007/s00335-002-4001-1.View ArticlePubMedGoogle Scholar
- Loring JF, Porter JG, Seilhammer J, Kaser MR, Wesselschmidt R: A gene expression profile of embryonic stem cells and embryonic stem cell-derived neurons. Restor Neurol Neurosci. 2001, 18 (2–3): 81-8.PubMedGoogle Scholar
- Sato N, Sanjuan IM, Heke M, Uchida M, Naef F, Brivanlou AH: Molecular signature of human embryonic stem cells and its comparison with the mouse. Dev Biol. 2003, 260: 404-13. 10.1016/S0012-1606(03)00256-2.View ArticlePubMedGoogle Scholar
- Abeyta MJ, Clark AT, Rodriguez RT, Bodnar MS, Pera RA, Firpo MT: Unique gene expression signatures of independently-derived human embryonic stem cell lines. Hum Mol Genet. 2004, 13: 601-8. 10.1093/hmg/ddh068.View ArticlePubMedGoogle Scholar
- Richards M, Tan SP, Tan JH, Chan WK, Bongso A: The transcriptome profile of human embryonic stem cells as defined by SAGE. Stem Cells. 2004, 22: 51-64.View ArticlePubMedGoogle Scholar
- Brandenberger R, Wei H, Zhang S, Lei S, Murage J, Fisk GJ, Li Y, Xu C, Fang R, Guegler K, et al: Transcriptome characterization elucidates signaling networks that control human ES cell growth and differentiation. Nat Biotechnol. 2004, 22: 707-16. 10.1038/nbt971.View ArticlePubMedGoogle Scholar
- Bhattacharya B, Miura T, Brandenberger R, Mejido J, Luo Y, Yang AX, Joshi BH, Ginis I, Thies RS, Amit M, et al: Gene expression in human embryonic stem cell lines: unique molecular signature. Blood. 2004, 103: 2956-64. 10.1182/blood-2003-09-3314.View ArticlePubMedGoogle Scholar
- Sperger JM, Chen X, Draper JS, Antosiewicz JE, Chon CH, Jones SB, Brooks JD, Andrews PW, Brown PO, Thomson JA: Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors. Proc Natl Acad Sci U S A. 2003, 100: 13350-5. 10.1073/pnas.2235735100.PubMed CentralView ArticlePubMedGoogle Scholar
- Zeng X, Miura T, Luo Y, Bhattacharya B, Condie B, Chen J, Ginis I, Lyons I, Mejido J, Puri RK, et al: Properties of pluripotent human embryonic stem cells BG01 and BG02. Stem Cells. 2004, 22: 292-312.View ArticlePubMedGoogle Scholar
- Pera MF, Reubinoff B, Trounson A: Human embryonic stem cells. J Cell Sci. 2000, 113: 5-10.PubMedGoogle Scholar
- Amit M, Margulets V, Segev H, Shariki K, Laevsky I, Coleman R, Itskovitz-Eldor J: Human feeder layers for human embryonic stem cells. Biol Reprod. 2003, 68: 2150-6.View ArticlePubMedGoogle Scholar
- Brenner S, Johnson M, Bridgham J, Golda G, Lloyd DH, Johnson D, Luo S, McCurdy S, Foy M, Ewan M, et al: Gene expression analysis by massively parallel signature sequencing (MPSS) on microbead arrays. Nat Biotechnol. 2000, 18: 630-4. 10.1038/76469.View ArticlePubMedGoogle Scholar
- Jongeneel CV, Iseli C, Stevenson BJ, Riggins GJ, Lal A, Mackay A, RA Harris, O'Hare MJ, Neville AM, Simpson AJ, et al: Comprehensive sampling of gene expression in human cell lines with massively parallel signature sequencing. Proc Natl Acad Sci U S A. 2003, 100: 4702-5. 10.1073/pnas.0831040100.PubMed CentralView ArticlePubMedGoogle Scholar
- Carpenter MK, Rosler ES, Fisk GJ, Brandenberger R, Ares X, Miura T, Lucero M, Rao MS: Properties of four human embryonic stem cell lines maintained in a feeder-free culture system. Dev Dyn. 2004, 229: 243-58. 10.1002/dvdy.10431.View ArticlePubMedGoogle Scholar
- Rosler ES, Fisk GJ, Ares X, Irving J, Miura T, Rao MS, Carpenter MK: Long-term culture of human embryonic stem cells in feeder-free conditions. Dev Dyn. 2004, 229: 259-74. 10.1002/dvdy.10430.View ArticlePubMedGoogle Scholar
- Xu C, Inokuma MS, Denham J, Golds K, Kundu P, Gold JD, Carpenter MK: Feeder-free growth of undifferentiated human embryonic stem cells. Nat Biotechnol. 2001, 19: 971-4. 10.1038/nbt1001-971.View ArticlePubMedGoogle Scholar
- Carpenter MK, Rosler E, Rao MS: Characterization and differentiation of human embryonic stem cells. Cloning Stem Cells. 2003, 5: 79-88. 10.1089/153623003321512193.View ArticlePubMedGoogle Scholar
- Amit M, Shariki C, Margulets V, Itskovitz-Eldor J: Feeder layer- and serum-free culture of human embryonic stem cells. Biol Reprod. 2004, 70: 837-45.View ArticlePubMedGoogle Scholar
- Takahashi K, Mitsui K, Yamanaka S: Role of ERas in promoting tumour-like properties in mouse embryonic stem cells. Nature. 2003, 423: 541-5. 10.1038/nature01646.View ArticlePubMedGoogle Scholar
- Zemel S, Bartolomei MS, Tilghman SM: Physical linkage of two mammalian imprinted genes, H19 and insulin-like growth factor 2. Nat Genet. 1992, 2: 61-5. 10.1038/ng0992-61.View ArticlePubMedGoogle Scholar
- Leighton PA, Saam JR, Ingram RS, Tilghman SM: Genomic imprinting in mice: its function and mechanism. Biol Reprod. 1996, 54: 273-8.View ArticlePubMedGoogle Scholar
- Brenner S, Williams SR, Vermaas EH, Storck T, Moon K, McCollum C, Mao JI, Luo S, Kirchner JJ, Eletr S, et al: In vitro cloning of complex mixtures of DNA on microbeads: physical separation of differentially expressed cDNAs. Proc Natl Acad Sci U S A. 2000, 97: 1665-70. 10.1073/pnas.97.4.1665.PubMed CentralView ArticlePubMedGoogle Scholar
- Siegel S, Castellan N: Nonparametric Statistics for the Behavioral Sciences. 1988, 2Google Scholar
This article is published under license to BioMed Central Ltd. This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.