Reorganization of nuclear lamina – genome interactions upon differentiation of embryonic stem cells.

ABSTRACT: The three-dimensional organization of chromosomes within the nucleus and its dynamics during differentiation are largely unknown. We present a genome-wide analysis of the interactions between chromatin and the nuclear lamina during differentiation of mouse embryonic stem cells (ESCs) into lineage-committed neural precursor cells (NPCs) and terminally differentiated astrocytes. Chromatin in each of these cell types shows a similar organization into large lamina associated domains (LADs), which represent a transcriptionally repressive environment. During sequential differentiation steps, lamina interactions are progressively modified at hundreds of genomic locations. This remodeling is typically confined to individual transcription units and involves many genes that determine cellular identity. From ESCs to NPCs, the majority of genes that move away from the lamina are concomitantly activated. Strikingly, a significant amount remain inactive yet become primed for activation by further differentiation. These results suggest that lamina-genome interactions are widely involved in the control of gene expression programs during lineage commitment and terminal differentiation. UPDATE WM20121119: We have added per-probe Hidden Markov Model state calls for the 4 cell types. These calls are not used in Peric-Hupkes, Meuleman et al. (2010), but are used in Meuleman, Peric-Hupkes et al. (2012). The procedure to arrive at this HMM state calls is as follows: We fitted a two-state HMM whereby emissions are distributed as Student's t variables. Mean and variance of DamID signals differ between states, but the degree of freedom (nu) is the same. Gaps in the probe coverage were filled by evenly spaced null probe-values. The parameters were estimated by an adaptation of the ECME algorithm to the HMM framework, showing faster convergence than regular EM when nu is unknown (Filion et al., Cell, 2010). State calls were derived through the Viterbi algorithm. This process was repeated separately for each cell type, yielding per-probe calls. Probes in the ‘bound’ (1) state are indicated as LAD-probes, probes in the ‘unbound’ (0) state as inter-LAD-probes. UPDATE WM20130301: We have also added (1-based) BED files of constitutive LAD and inter-LAD regions (mm9) across the 4 cell types. These regions are the result of simply concatenating directly adjacent cLAD/ciLAD Hidden Markov Model state calls. Again, these calls are not used in Peric-Hupkes, Meuleman et al. (2010), but are used in Meuleman, Peric-Hupkes et al. (2012).

ORGANISM(S): Mus musculus

PROVIDER: GSE17051 | GEO | 2010/05/27

SECONDARY ACCESSION(S): PRJNA117779

REPOSITORIES: GEO