CoChat

Genetic Background Drives Transcriptional Variation in Human Induced Pluripotent Stem Cells

Open access
Rouhani, Foad, Kumasaka, Natsuhiko, de Brito, Miguel Cardoso, Bradley, Allan, Vallier, Ludovic, Gaffney, Daniel · 2014 · PLoS Genetics

The DOI matches its Crossref record; 35 of the 35 references it cites were found in a registry or an index.

Abstract

Human iPS cells have been generated using a diverse range of tissues from a variety of donors using different reprogramming vectors. However, these cell lines are heterogeneous, which presents a limitation for their use in disease modeling and personalized medicine. To explore the basis of this heterogeneity we generated 25 iPS cell lines under normalised conditions from the same set of somatic tissues across a number of donors. RNA-seq data sets from each cell line were compared to identify the majority contributors to transcriptional heterogeneity. We found that genetic differences between individual donors were the major cause of transcriptional variation between lines. In contrast, residual signatures from the somatic cell of origin, so called epigenetic memory, contributed relatively little to transcriptional variation. Thus, underlying genetic background variation is responsible for most heterogeneity between human iPS cell lines. We conclude that epigenetic effects in hIPSCs are minimal, and that hIPSCs are a stable, robust and powerful platform for large-scale studies of the function of genetic differences between individuals. Our data also suggest that future studies using hIPSCs as a model system should focus most effort on collection of large numbers of donors, rather than generating large numbers of lines from the same donor.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1371/journal.pgen.1004432

Cites

35 of 35 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Experimental design and variance components analysis of transcription in iPSCs, somatic progenitors, H9 and Val9 embryonic stem cells. (a) Schematic showing experimental design. (b) Heatmap of Pearson correlation coefficients of log2 FPKMs (fragments per kilobase of exon per million fragments mapped) across all genes in all samples (N = 47). Complete-linkage clustering on the correlation coefficients was performed to order samples by similarity. (c) Decomposition of transcriptional variation using a linear mixed model. Each bar shows the percentage of transcriptional variance explained (%VE) by each of the components, holding all other factors constant (see Text S1 section 1.3 for details). Bars show the percentage variance explained between iPS and ES cells δ2 22/(δ2 22+σ2) (orange); between adult tissues δ2 21/(δ2 21+σ2) (purple); between iPSC somatic tissue of origin δ2 32/(δ2 32+σ2) (pink); between individual (iPSCs) δ2 41/(δ2 42+σ2) (red); between individual (adult tissues) δ2 41/(δ2 41+σ2) (green) and between sequencing batches δ2 5/(δ2 5+σ2)(blue). δ2 and σ2 parameters are explained in section 1.3 of Text S1. Red text denotes variance explained by a component in a model that excluded an individual component, black text denotes variance explained in a model including an individual component. Panels show layered covariance matrices Zj Dj Zj for the variance components j (j = 2…5) of which the correlation heatmap in Figure 1b consists. The color corresponds to each of variance components in the barchart of Figure 1c and white color corresponds to covariance of 0. The sample order of the layered matrices is compatible with Figure 1b. (d) Bars show the estimate of residual variance parameters estimated from a homoscedastic and two heteroscedastic models. “Homoscedastic model” represents a model with a single error term shared over all samples in the data set, and the estimate of this error term is labeled “All”. The heteroscedastic models 1 and 2 are described in the main text and in the Supplement (Section 1.3.3. “Heteroscedastic Model”. Each bar shows the estimate of a residual error parameter as outlined the Supplement section 1.3.3.
Figure 1 Experimental design and variance components analysis of transcription in iPSCs, somatic progenitors, H9 and Val9 embryonic stem cells. (a) Schematic showing experimental design. (b) Heatmap of Pearson correlation coefficients of log2 FPKMs (fragments per kilobase of exon per million fragments mapped) across all genes in all samples (N = 47). Complete-linkage clustering on the correlation coefficients was performed to order samples by similarity. (c) Decomposition of transcriptional variation using a linear mixed model. Each bar shows the percentage of transcriptional variance explained (%VE) by each of the components, holding all other factors constant (see Text S1 section 1.3 for details). Bars show the percentage variance explained between iPS and ES cells δ2 22/(δ2 22+σ2) (orange); between adult tissues δ2 21/(δ2 21+σ2) (purple); between iPSC somatic tissue of origin δ2 32/(δ2 32+σ2) (pink); between individual (iPSCs) δ2 41/(δ2 42+σ2) (red); between individual (adult tissues) δ2 41/(δ2 41+σ2) (green) and between sequencing batches δ2 5/(δ2 5+σ2)(blue). δ2 and σ2 parameters are explained in section 1.3 of Text S1. Red text denotes variance explained by a component in a model that excluded an individual component, black text denotes variance explained in a model including an individual component. Panels show layered covariance matrices Zj Dj Zj for the variance components j (j = 2…5) of which the correlation heatmap in Figure 1b consists. The color corresponds to each of variance components in the barchart of Figure 1c and white color corresponds to covariance of 0. The sample order of the layered matrices is compatible with Figure 1b. (d) Bars show the estimate of residual variance parameters estimated from a homoscedastic and two heteroscedastic models. “Homoscedastic model” represents a model with a single error term shared over all samples in the data set, and the estimate of this error term is labeled “All”. The heteroscedastic models 1 and 2 are described in the main text and in the Supplement (Section 1.3.3. “Heteroscedastic Model”. Each bar shows the estimate of a residual error parameter as outlined the Supplement section 1.3.3.CC BY · PubMed Central
Figure 2 Effects of cell type and tissues of origin on expression levels in iPSCs. (a) Schematic of hypothesis testing approach to identify genes with invariant expression (IE) from genes that we defined as correctly reprogrammed (CR), aberrantly reprogrammed (AR), with transcriptional memory (TM) and showing complex expression. Genes showing complex expression are further classified into partial AR (PAR) and partial TM (PTM) according to the expression patterns. (b) Percentages of genes assigned to alternative differential expression groups based on a hierarchical model (see Text S1 1.4.3). (c) Volcano plots of differentially expressed genes. Plots show log10 of minimum P-values among the four alternative hypotheses against log2 fold change of average expression levels between iPSCs and ESCs. The colours of points indicate the differential expression categories into which a gene was classified. Dashed lines show twofold enrichment of mean expression levels between iPSCs and ESCs. FDR thresholds were calculated by permutation. (d) Read coverage depth in three examples of differentially expressed genes; SOX2, was categorised as correctly reprogrammed in all three tissues (F-iPSCs: p<1×10−24; K-iPSCs: p<1.2×10−16;E-iPSCs: p<4.6×10−25), H19 was categorised as aberrantly reprogrammed in all three tissues (F-iPSCs: p<1.2×10−6; K-iPSCs: p<2.7×10−4;E-iPSCs: p<1.5×10−16) and TYW3 gene was categorised as partial transcriptional memory in E-iPSCs (p<7.1×10−8), invariant expression in F-iPSCs (p>0.1) and correctly reprogrammed in K-iPSCs (p<1.4×10−3). Coverage depth was truncated at 1000 reads per bp.
Figure 2 Effects of cell type and tissues of origin on expression levels in iPSCs. (a) Schematic of hypothesis testing approach to identify genes with invariant expression (IE) from genes that we defined as correctly reprogrammed (CR), aberrantly reprogrammed (AR), with transcriptional memory (TM) and showing complex expression. Genes showing complex expression are further classified into partial AR (PAR) and partial TM (PTM) according to the expression patterns. (b) Percentages of genes assigned to alternative differential expression groups based on a hierarchical model (see Text S1 1.4.3). (c) Volcano plots of differentially expressed genes. Plots show log10 of minimum P-values among the four alternative hypotheses against log2 fold change of average expression levels between iPSCs and ESCs. The colours of points indicate the differential expression categories into which a gene was classified. Dashed lines show twofold enrichment of mean expression levels between iPSCs and ESCs. FDR thresholds were calculated by permutation. (d) Read coverage depth in three examples of differentially expressed genes; SOX2, was categorised as correctly reprogrammed in all three tissues (F-iPSCs: p<1×10−24; K-iPSCs: p<1.2×10−16;E-iPSCs: p<4.6×10−25), H19 was categorised as aberrantly reprogrammed in all three tissues (F-iPSCs: p<1.2×10−6; K-iPSCs: p<2.7×10−4;E-iPSCs: p<1.5×10−16) and TYW3 gene was categorised as partial transcriptional memory in E-iPSCs (p<7.1×10−8), invariant expression in F-iPSCs (p>0.1) and correctly reprogrammed in K-iPSCs (p<1.4×10−3). Coverage depth was truncated at 1000 reads per bp.CC BY · PubMed Central
Figure 3 Imprinting effects on transcription in adult and iPS cells. (a) Allelic imbalance between adult cells and iPS cells for each donor (S2, S4 and S7). (b) The mosaic plots for each cell line show the allelic imbalance at multiple heterozygous SNP loci throughout the transcribed region. Each row corresponds to a sample, and each box corresponds to a single heterozygous SNP. Box width is proportional to the total nucleotide count at the heterozygous SNP, with the box bisected in proportion to the paternal and maternal nucleotide counts (blue and orange) based on the phasing haplotype information. Examples show consistent loss of imprinting in iPSCs (ZDBF2) and in adult cells (IFG2).
Figure 3 Imprinting effects on transcription in adult and iPS cells. (a) Allelic imbalance between adult cells and iPS cells for each donor (S2, S4 and S7). (b) The mosaic plots for each cell line show the allelic imbalance at multiple heterozygous SNP loci throughout the transcribed region. Each row corresponds to a sample, and each box corresponds to a single heterozygous SNP. Box width is proportional to the total nucleotide count at the heterozygous SNP, with the box bisected in proportion to the paternal and maternal nucleotide counts (blue and orange) based on the phasing haplotype information. Examples show consistent loss of imprinting in iPSCs (ZDBF2) and in adult cells (IFG2).CC BY · PubMed Central
Figure 4 Genetic effects on transcription in iPSCs. (a) Variance component of known eQTL genes showing primary eQTL SNPs being able to explain 17% of FPKM variation on average. Top and bottom 3,000 eQTL genes and SNPs were determined by means of P-values using gEUVADIS RNA-seq data [27]. (b) Box- plot of log2 FPKM aggregated across 462 eQTL genes (FDR 5%) stratified into high (+/+), medium (+/−) and low (−/−) genotypes at the minimum p-value eQTL SNPs ascertained from [2] (FDR 5%). Pearson correlation r = 0.44, p<9.8×10−68. (c) Boxplot of allelic imbalance at high-expression (+) and low-expression (−) haplotypes across all eQTL genes ascertained from [2]. The red line is at 0.5 indicating the expected fraction under the null. Student t p<1.3×10−8. (d) Mean RNA-Seq coverage depth in iPSCs, averaged across all iPSC lines from each individual, at a putative example eQTL for the noncoding RNA SNHG5 gene. Individuals are grouped by their genotype at the putatively causal variant rs1059307. Individual genotypes are shown with the number of RNA-seq samples in parentheses. (e) Average coverage depth of fragments coming from the high-expression haplotype (allele T) and low- expression haplotype (allele G) at rs1059307. (f) eQTL association of SNHG5 gene at rs1059307 in lymphoblastoid cell lines from [2] showing distributions of log2 FPKM against SNP genotypes at rs1059307. The red line shows the best-fit linear regression line.
Figure 4 Genetic effects on transcription in iPSCs. (a) Variance component of known eQTL genes showing primary eQTL SNPs being able to explain 17% of FPKM variation on average. Top and bottom 3,000 eQTL genes and SNPs were determined by means of P-values using gEUVADIS RNA-seq data [27]. (b) Box- plot of log2 FPKM aggregated across 462 eQTL genes (FDR 5%) stratified into high (+/+), medium (+/−) and low (−/−) genotypes at the minimum p-value eQTL SNPs ascertained from [2] (FDR 5%). Pearson correlation r = 0.44, p<9.8×10−68. (c) Boxplot of allelic imbalance at high-expression (+) and low-expression (−) haplotypes across all eQTL genes ascertained from [2]. The red line is at 0.5 indicating the expected fraction under the null. Student t p<1.3×10−8. (d) Mean RNA-Seq coverage depth in iPSCs, averaged across all iPSC lines from each individual, at a putative example eQTL for the noncoding RNA SNHG5 gene. Individuals are grouped by their genotype at the putatively causal variant rs1059307. Individual genotypes are shown with the number of RNA-seq samples in parentheses. (e) Average coverage depth of fragments coming from the high-expression haplotype (allele T) and low- expression haplotype (allele G) at rs1059307. (f) eQTL association of SNHG5 gene at rs1059307 in lymphoblastoid cell lines from [2] showing distributions of log2 FPKM against SNP genotypes at rs1059307. The red line shows the best-fit linear regression line.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 35 references

Cited by

Items 201–299 of 299

Metadata

Authors
Rouhani, F., Kumasaka, N., de Brito, M. C., Bradley, A., Vallier, L., & Gaffney, D.
Affiliations
  • Wellcome Sanger Institute (GB)
  • University of Cambridge (GB)
Corresponding author
Gaffney, Daniel
Venue
PLoS Genetics
Published
2014-06-05
Volume
10
Issue
6
Pages
e1004432
ISSN
  • 1553-7390 (ISSN-L)
  • 1553-7404
OpenAlex
W2008485310
Cited by
294
References
35
Citations by year
2026: 7 · 2025: 15 · 2024: 22 · 2023: 22 · 2022: 19 · 2021: 29 · 2020: 25 · 2019: 21 · 2018: 33 · 2017: 37 · 2016: 33 · 2015: 25 · 2014: 5
FWCI
13.59
Type
journal article
Language
en
Open access
Locations
  • PLoS Genetics (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • Figshare (submittedVersion, cc-by)
  • Apollo (University of Cambridge) (submittedVersion, cc-by)
  • Apollo (cc-by)
Funders
  • Wellcome Trust (EU-FP7-, 098051)
  • NIHR Biomedical Research Centre, Royal Marsden NHS Foundation Trust/Institute of Cancer Research
  • National Institute for Health and Care Research
  • Engineering and Physical Sciences Research Council (TS/H001220/1)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Pluripotent Stem Cells Research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Epigenetics and DNA Methylation
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Developmental Biology and Gene Regulation
Keywords
human induced pluripotent stem cells, transcriptional heterogeneity, epigenetic memory, cell line heterogeneity, disease modeling, personalized medicine
MeSH terms
Adult, Base Sequence, Cell Differentiation, Cells, Cultured, Female, Fibroblasts, Humans, Male, Transcription, Genetic, Keratinocytes, Sequence Alignment, Sequence Analysis, RNA, Endothelial Cells, Embryonic Stem Cells, Induced Pluripotent Stem Cells, Epigenomics

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.