2.1. Gene expression data and processing

Single myofiber gene expression data were collected from Gene Expression Omnibus (GEO) and Sequence Read Archive (SRA) databases using the following IDs: {"type":"entrez-geo","attrs":{"text":"GSE98328","term_id":"98328"}}GSE98328, SRX2768351, SRX2768352, SRX2768353 [5], and {"type":"entrez-geo","attrs":{"text":"GSE112716","term_id":"112716"}}GSE112716 [12]. For snRNA-seq, we used processed data retrieved from [24], [25], [26] as an example of muscle pathology, fiber typing, and ageing respectively. Microarray gene expression data were processed as follow. Agilent microarray mouse platform was re-annotated (Gencode annotation release vM22, evidence-based annotation of the mouse genome GRCm38, version M22 Ensembl 97) both for coding and non-coding RNAs. Microarray data were normalized using quantile normalization separately for protein-coding and long non-coding genes. The dataset includes 10 biological replicates for each myofiber type considered (1, 2A, 2A/2X, 2X, 2X/2B, and 2B). Myofibers were sub-grouped in transcriptional slow (type 1), transcriptional intermediate (type 2A, 2A/2X, and 2X), and transcriptional fast (type 2X/2B and 2B). RNA sequencing data for miRNA identification were mapped to the known mouse miRNA precursors from the miRBase database (Ver. 19) using the mapper module of miRDeep with default settings. Quantize module was used to normalize read counts of mature miRNAs.