Figure 1

NMJ-Analyser workflow. NMJ-Analyser consists of four sequential steps. (Step 1) Muscles are dissected, immunostained and digitalised (here, hindlimb lumbrical muscles). NMJ innervation status is manually assessed (full, partial, denervated) using an imaging package. (Step 2) Morphological analysis of NMJs requires raw images (plane view) to be formatted (.TIFF/.JPEG/.PNG). Features are captured for individual pre- or post-synaptic NMJs and for parameters involving interactions between these two components. (Step 3) Data from NMJ innervation status and morphological features are integrated into a matrix. Additional manual curation can be performed to confirm the whole individual NMJ was captured and its entire 3D structure analysed using NMJ-Analyser. (Step 4) Curated data output are divided into training and predictive sets. The training set is input into a machine learning model while the predictive set is used to predict the NMJ innervation status. Additional File 2 contains a detailed tutorial of how to use NMJ-Analyser.