Limits...
Comparative computational analysis of pluripotency in human and mouse stem cells.

Ernst M, Abu Dawud R, Kurtz A, Schotta G, Taher L, Fuellen G - Sci Rep (2015)

Bottom Line: Pluripotent cells can be subdivided into two distinct states, the naïve and the primed state, the latter being further advanced on the path of differentiation.Reprogramming of human stem cells into a more naïve-like state is an important research focus.The pipeline consists of identifying regulated start-ups/shut-downs in terms of molecular interactions, followed by functional annotation of the genes involved and aggregation of results across conditions, yielding sets of mechanisms that are consistently regulated in transitions towards similar states of pluripotency.

View Article: PubMed Central - PubMed

Affiliation: Institute for Biostatistics and Informatics in Medicine and Ageing Research, Rostock University Medical Center, Rostock, Germany.

ABSTRACT
Pluripotent cells can be subdivided into two distinct states, the naïve and the primed state, the latter being further advanced on the path of differentiation. There are substantial differences in the regulation of pluripotency between human and mouse, and in humans only stem cells that resemble the primed state in mouse are readily available. Reprogramming of human stem cells into a more naïve-like state is an important research focus. Here, we developed a pipeline to reanalyze transcriptomics data sets that describe both states, naïve and primed pluripotency, in human and mouse. The pipeline consists of identifying regulated start-ups/shut-downs in terms of molecular interactions, followed by functional annotation of the genes involved and aggregation of results across conditions, yielding sets of mechanisms that are consistently regulated in transitions towards similar states of pluripotency. Our results suggest that one published protocol for naïve human cells gave rise to human cells that indeed share putative mechanisms with the prototypical naïve mouse pluripotent cells, such as DNA damage response and histone acetylation. However, cellular response and differentiation-related mechanisms are similar between the naïve human state and the primed mouse state, so the naïve human state did not fully reflect the naïve mouse state.

Show MeSH

Related in: MedlinePlus

Principal component analysis (PCA) of evidence for enrichment of GOBP terms in target aggregates.Aggregation of the twelve comparisons into target aggregates was performed as described in the text; the matrix of log-transformed p-values (Figure 6) was then analysed by PCA. In this biplot, two scatterplots are overlaid: one for the scores of the four target aggregates, represented by their respective abbreviations (see Figure 3, denoted NH, PH, NM, and PM, for the four pairs of a species and a state), and one for the relative importance (or loadings) of the GOBP terms, plotted as grey or coloured dot symbols. The symbol colour indicates for which, if any, block aggregate (as defined in Figure 3, denoted N, P, NHPM, and NMPH) the term was found to be significant (see text). The axes below and to the left of the plot belong to the scores plot, the ones on top and to the right to the loadings plot.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4300513&req=5

f7: Principal component analysis (PCA) of evidence for enrichment of GOBP terms in target aggregates.Aggregation of the twelve comparisons into target aggregates was performed as described in the text; the matrix of log-transformed p-values (Figure 6) was then analysed by PCA. In this biplot, two scatterplots are overlaid: one for the scores of the four target aggregates, represented by their respective abbreviations (see Figure 3, denoted NH, PH, NM, and PM, for the four pairs of a species and a state), and one for the relative importance (or loadings) of the GOBP terms, plotted as grey or coloured dot symbols. The symbol colour indicates for which, if any, block aggregate (as defined in Figure 3, denoted N, P, NHPM, and NMPH) the term was found to be significant (see text). The axes below and to the left of the plot belong to the scores plot, the ones on top and to the right to the loadings plot.

Mentions: Applying a PCA to the matrix of aggregated evidence highlights the relationships between the target aggregates as well as the relative importance (loadings) of the single GOBP terms for the target aggregates (Figure 7). Here, the first PC clearly separates the target aggregates PM and PH (which have positive scores) from the target aggregates NM and NH (negative scores), with both groups having a lower intragroup than intergroup distance. Again, we conclude that there are functional similarities between the naïve states in mouse (represented by the target aggregate NM) and human (NH). Moreover, since the first PC accounts for 74% of the variance, these similarities are substantial. However, Figure 4 indicated that the naïve block aggregate is characterized by depletion of GOBP terms that are enriched elsewhere; yet the aggregation procedure that underlies the PCA considers enriched GOBP terms only. We investigated this issue further using a statistical approach, determining how many and which GOBP terms and, hence, biological processes are in fact common between NM and NH (see Figure 8 below). While the first PC in Figure 7 seems to capture functional and species-independent characteristics of naïve and primed pluripotency, the second PC, explaining roughly 17% of total variance, has a different interpretation. Here, the target aggregates NH and PM have strikingly similar scores, hinting at some biological process similarity between the claimed naïve state in human and the primed state in mouse. This finding prompted us to define the (additional) block aggregate NHPM (see Figure 3). In the next section we will use this aggregate to decipher the GOBP terms that might underlie this finding and that calls for improvements of protocols for human naïvity.


Comparative computational analysis of pluripotency in human and mouse stem cells.

Ernst M, Abu Dawud R, Kurtz A, Schotta G, Taher L, Fuellen G - Sci Rep (2015)

Principal component analysis (PCA) of evidence for enrichment of GOBP terms in target aggregates.Aggregation of the twelve comparisons into target aggregates was performed as described in the text; the matrix of log-transformed p-values (Figure 6) was then analysed by PCA. In this biplot, two scatterplots are overlaid: one for the scores of the four target aggregates, represented by their respective abbreviations (see Figure 3, denoted NH, PH, NM, and PM, for the four pairs of a species and a state), and one for the relative importance (or loadings) of the GOBP terms, plotted as grey or coloured dot symbols. The symbol colour indicates for which, if any, block aggregate (as defined in Figure 3, denoted N, P, NHPM, and NMPH) the term was found to be significant (see text). The axes below and to the left of the plot belong to the scores plot, the ones on top and to the right to the loadings plot.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4300513&req=5

f7: Principal component analysis (PCA) of evidence for enrichment of GOBP terms in target aggregates.Aggregation of the twelve comparisons into target aggregates was performed as described in the text; the matrix of log-transformed p-values (Figure 6) was then analysed by PCA. In this biplot, two scatterplots are overlaid: one for the scores of the four target aggregates, represented by their respective abbreviations (see Figure 3, denoted NH, PH, NM, and PM, for the four pairs of a species and a state), and one for the relative importance (or loadings) of the GOBP terms, plotted as grey or coloured dot symbols. The symbol colour indicates for which, if any, block aggregate (as defined in Figure 3, denoted N, P, NHPM, and NMPH) the term was found to be significant (see text). The axes below and to the left of the plot belong to the scores plot, the ones on top and to the right to the loadings plot.
Mentions: Applying a PCA to the matrix of aggregated evidence highlights the relationships between the target aggregates as well as the relative importance (loadings) of the single GOBP terms for the target aggregates (Figure 7). Here, the first PC clearly separates the target aggregates PM and PH (which have positive scores) from the target aggregates NM and NH (negative scores), with both groups having a lower intragroup than intergroup distance. Again, we conclude that there are functional similarities between the naïve states in mouse (represented by the target aggregate NM) and human (NH). Moreover, since the first PC accounts for 74% of the variance, these similarities are substantial. However, Figure 4 indicated that the naïve block aggregate is characterized by depletion of GOBP terms that are enriched elsewhere; yet the aggregation procedure that underlies the PCA considers enriched GOBP terms only. We investigated this issue further using a statistical approach, determining how many and which GOBP terms and, hence, biological processes are in fact common between NM and NH (see Figure 8 below). While the first PC in Figure 7 seems to capture functional and species-independent characteristics of naïve and primed pluripotency, the second PC, explaining roughly 17% of total variance, has a different interpretation. Here, the target aggregates NH and PM have strikingly similar scores, hinting at some biological process similarity between the claimed naïve state in human and the primed state in mouse. This finding prompted us to define the (additional) block aggregate NHPM (see Figure 3). In the next section we will use this aggregate to decipher the GOBP terms that might underlie this finding and that calls for improvements of protocols for human naïvity.

Bottom Line: Pluripotent cells can be subdivided into two distinct states, the naïve and the primed state, the latter being further advanced on the path of differentiation.Reprogramming of human stem cells into a more naïve-like state is an important research focus.The pipeline consists of identifying regulated start-ups/shut-downs in terms of molecular interactions, followed by functional annotation of the genes involved and aggregation of results across conditions, yielding sets of mechanisms that are consistently regulated in transitions towards similar states of pluripotency.

View Article: PubMed Central - PubMed

Affiliation: Institute for Biostatistics and Informatics in Medicine and Ageing Research, Rostock University Medical Center, Rostock, Germany.

ABSTRACT
Pluripotent cells can be subdivided into two distinct states, the naïve and the primed state, the latter being further advanced on the path of differentiation. There are substantial differences in the regulation of pluripotency between human and mouse, and in humans only stem cells that resemble the primed state in mouse are readily available. Reprogramming of human stem cells into a more naïve-like state is an important research focus. Here, we developed a pipeline to reanalyze transcriptomics data sets that describe both states, naïve and primed pluripotency, in human and mouse. The pipeline consists of identifying regulated start-ups/shut-downs in terms of molecular interactions, followed by functional annotation of the genes involved and aggregation of results across conditions, yielding sets of mechanisms that are consistently regulated in transitions towards similar states of pluripotency. Our results suggest that one published protocol for naïve human cells gave rise to human cells that indeed share putative mechanisms with the prototypical naïve mouse pluripotent cells, such as DNA damage response and histone acetylation. However, cellular response and differentiation-related mechanisms are similar between the naïve human state and the primed mouse state, so the naïve human state did not fully reflect the naïve mouse state.

Show MeSH
Related in: MedlinePlus