Fig. 3: Clinical features and transcriptional spectrum.

The spectrum order was validated using both in-silico algorithms and clinical features from different myopathies. Given that there are no linear relationships in the muscle UMAP plot, for each myopathy, we categorized the muscles into three categories by their UMAP X location and tested whether an ordered relationship exists among these three groups using the Jonckheere-Terpstra test. A The results of the in-silico pseudo-time and trajectory algorithms were consistent with the GTEx inference. B CTG expansions (in peripheral blood) increased with the spectrum progression in congenital myotonic dystrophy (CDM) patients. C Mercuri scores, as well as 10-meter and 6-minute walk test results, were also consistent with disease progression in limb girdle muscular dystrophies R12 (LGMD R12) patients. The more compact distribution seen in LGMD R12, as compared to CDM and FSHD, might be due to the repeated biopsies taken. For each LGMD R12 patient, three biopsies were collected (represented by three purple dots), whereas for each CDM and FSHD patient, only one biopsy was taken (represented by a single purple dot). D Fat fractions, pathology, and clinical severity scores of facioscapulohumeral muscular dystrophy (FSHD) patients did not show as clear a trend with disease progression as seen in CDM and LGMD R12. This is presumed to be related to the heterogeneous muscle locations (thigh and leg) in the FSHD study.