Figure 2. Modular epigenome model to explain in vivo rejuvenation results. A bimodular epigenome is considered, where Module A represents the DNAm clock component which encompasses all age-dependent DNA methylation epigenetic marks. Module B represents the remainder of the epigenome, including but not limited to cell identity marks. The model does not make further assumptions on the properties of each module. In this context, the existence of, for instance, age-dependent cell identity marks (bivalent marks) is not ruled out. The upper diagram (path 1) proposes that ex vivo, conventional reprogramming erases all age and cell identity marks from both modules, and can turn an old (brown color represents old) neuron or any other cell, into an iPSC that can then be differentiated back to a rejuvenated (blue represents young) neuron. In vivo, continuous expression of the OSKM genes leads to the genesis of multiple teratomas and the death of the animal. The middle diagram (path 2) illustrates the hypothesis that several cycles of partial reprogramming can progressively rejuvenate cells by erasing all epigenetic marks of age without affecting cell type identity marks. This means that in principle, the strategy could lead to major phenotype rejuvenation in vivo. The lower diagram (path 3) illustrates an alternative outcome for partial reprogramming. In this case, partial reprogramming erases age marks from the DNAm clock (module A) but spares age marks of module B. This outcome is compatible with a major resetting of DNAm age but partial rejuvenation of the phenotype.