Acetoacetate (AA) is a ketone body and acts as a fuel to supply energy for cellular activity of various tissues. Here, we uncovered a novel function of AA in promoting muscle cell proliferation. Notably, the functional role of AA in regulating muscle cell function is further evidenced by its capability to accelerate muscle regeneration in normal mice, and it ameliorates muscular dystrophy in mdx mice. Mechanistically, our data from multiparameter analyses consistently support the notion that AA plays a non-metabolic role in regulating muscle cell function. Finally, we show that AA exerts its function through activation of the MEK1-ERK1/2-cyclin D1 pathway, revealing a novel mechanism in which AA serves as a signaling metabolite in mediating muscle cell function. Our findings highlight the profound functions of a small metabolite as signaling molecule in mammalian cells.Satellite cells, which are among the most abundant well defined adult stem cell types in skeletal muscle, play functionally important roles in postnatal growth, repair, and the regeneration of skeletal muscle (1-6). Because of their powerful ability to regenerate in vivo in response to muscle damage and various stimuli, satellite cells represent important targets for the treatment of muscular diseases (7-10). The recent development of stem cell-based regenerative medicine strategies has brought enormous interest in the discovery of regulatory factors capable of controlling satellite cell functions, such as activation, proliferation, differentiation, and self-renewal (11-13). Identification of such factors is expected to not only improve our understanding of the regulatory mechanisms that govern satellite cell functions, but also to facilitate the development of stem cell-based therapies for the treatment of muscular dystrophy or other chronic diseases associated with muscle wasting.Recent studies demonstrating a close correlation between cell proliferation and metabolic alterations in various tumor types have drawn attention to the significance of intrinsic small metabolites as signaling molecules responsible for regulating various cellular activities (14, 15). Although only a very limited number of such metabolites have been identified to date, accumulating evidence suggests that these metabolites can be oncogenic and alter cell signaling through epigenetic regulation. For example, 2-hydroxyglutarate (2-HG), 4 succinate, and fumarate, which are the best characterized small metabolites with oncogenic function, have come to be regarded as oncometabolites (16 -19). In tumor cells, 2-HG is generated by mutant forms of isocitrate dehydrogenase (IDH1 and IDH2) (20 -23), whereas succinate and fumarate accumulate via mutant forms of succinate dehydrogenase and fumarate hydratase, respectively (24 -27). It has been clearly demonstrated that increases in the levels of these oncometabolites play causal roles in tumorigenesis (26 -34). Recent studies of the molecular mechanisms underlying their action have revealed that 2-HG and elevated levels of succinate ...

The latest progresses of experimental biology have generated a large number of data with different formats and lengths. Deep learning is an ideal tool to deal with complex datasets, but its inherent “black box” nature needs more interpretability. At the same time, traditional interpretable machine learning methods, such as linear regression or random forest, could only deal with numerical features instead of modular features often encountered in the biological field. Here, we present MultiCapsNet (https://github.com/wanglf19/MultiCapsNet), a new deep learning model built on CapsNet and scCapsNet, which possesses the merits such as easy data integration and high model interpretability. To demonstrate the ability of this model as an interpretable classifier to deal with modular inputs, we test MultiCapsNet on three datasets with different data type and application scenarios. Firstly, on the labeled variant call dataset, MultiCapsNet shows a similar classification performance with neural network model, and provides importance scores for data sources directly without an extra importance determination step required by the neural network model. The importance scores generated by these two models are highly correlated. Secondly, on single cell RNA sequence (scRNA-seq) dataset, MultiCapsNet integrates information about protein-protein interaction (PPI), and protein-DNA interaction (PDI). The classification accuracy of MultiCapsNet is comparable to the neural network and random forest model. Meanwhile, MultiCapsNet reveals how each transcription factor (TF) or PPI cluster node contributes to classification of cell type. Thirdly, we made a comparison between MultiCapsNet and SCENIC. The results show several cell type relevant TFs identified by both methods, further proving the validity and interpretability of the MultiCapsNet.