Skip to main content
. 2016 Feb 22;5:e11469. doi: 10.7554/eLife.11469

Figure 5. The DBN recapitulates the consequences of TAL1 and LYL1 single and double perturbations as seen in vivo and in vitro.

Computational prediction of gene expression patterns for the nine TFs of interest after perturbation of TAL1 (a), LYL1 (b) or both (c). Deletion of TAL1 or LYL1 on their own has no major consequences on the expression levels of the other eight TFs of the gene regulatory network, but simultaneous deletion of both TAL1 and LYL1 caused changes in expression of several genes, mainly a decrease in Gata2 and Runx1. This major disruption of the core GRN for blood stem/progenitor cells is therefore consistent with TAL1/LYL1 double knockout HSCs showing a much more severe phenotype than the respective single knock-outs. One thousand simulations were run for each perturbation to determine the TFs expression levels in a 'cell population' by selecting expression levels at random time points after reaching its initial steady state. Expression levels of 0 resemble no expression, whereas expression levels of 1 stand for highest expression level that is possible in this system. The scale of the x-axes is linear. (d) Gene expression levels measured in single 416b cells transfected with siRNA constructs against Tal1 or a control. The density plots of gene expression levels after perturbation of TAL1 indicate the relative number of cells (y-axes) at each expression level (x-axes). The scale of the x-axes is linear. The values indicate the results of the Wilcoxon rank-sum test: alterations to the expression profiles are indicated by the p-value (statistical significance: p <0.001 for computational data and p <0.05 for experimental data); substantial shifts in median expression level are indicated by the shift of median (SOM) (SOM >0.1 for computational data and >1 for experimental data). For details, see Figure 5—figure supplement 1; for full expression data, see Figure 5—source data 1 .

DOI: http://dx.doi.org/10.7554/eLife.11469.046

Figure 5—source data 1. Raw and normalised data for the single cell gene expression experiments presented in this study.
1) TAL1 down-regulation (related to Figure 5 d), 2) PU.1 down-regulation (related to Figure 6 a), 3) GFI1B up-regulation (related to Figure 6b) and 4) AML-ETO9a perturbation (related to Figure 6 c)
elife-11469-fig5-data1.xlsx (250.8KB, xlsx)
DOI: 10.7554/eLife.11469.047

Figure 5.

Figure 5—figure supplement 1. Significance tests for the computational and experimental data after TF perturbations.

Figure 5—figure supplement 1.

To determine statistical significance the Wilcoxon rank-sum test was used. Alterations to the expression profiles are indicated by the p-value; with statistically significance defined as follows: p <0.001 for computational data and p <0.05 for experimental data. Significance of a substantial shift in median expression levels are as follows: shift of median >0.1 for computational data and >1 for experimental data (because of different scales). If the number for the shift of median is negative, the median of the perturbation data is smaller than that of the wild-type control; if the number is positive, the median of the perturbation is larger than that of the control. For simplicity, all significant changes are highlighted in red (p-value) and blue (shift of median).
Figure 5—figure supplement 2. Histogram plots showing the gene expression distributions of all nine genes of the network for the perturbations presented in this study.

Figure 5—figure supplement 2.

(a) LYL1 down-regulation; (b) TAL1/SCL down-regulation; (c) LYL1 and TAL1/SCL down-regulation; (d) PU.1 down-regulation; (e) GFI1B up-regulation; and (f) AML-ETO9a simulation.