Muscle function within an organism depends on the feedback between molecular and meter-scale processes. Although the motions of muscle’s contractile machinery are well described in isolated preparations, only a handful of experiments have documented the kinematics of the lattice occurring when multi-scale interactions are fully intact. We used time-resolved X-ray diffraction to record the kinematics of the myofilament lattice within a normal operating context: the tethered flight of Manduca sexta. As the primary flight muscles of M. sexta are synchronous, we used these results to reveal the timing of in vivo cross-bridge recruitment, which occurred 24 ms (s.d. 26) following activation. In addition, the thick filaments stretched an average of 0.75% (s.d. 0.32) and thin filaments stretched 1.11% (s.d. 0.65). In contrast to other in vivo preparations, lattice spacing changed an average of 2.72% (s.d. 1.47). Lattice dilation of this magnitude significantly affects shortening velocity and force generation, and filament stretching tunes force generation. While the kinematics were consistent within individual trials, there was extensive variation between trials. Using a mechanism-free machine learning model we searched for patterns within and across trials. Although lattice kinematics were predictable within trials, the model could not create predictions across trials. This indicates that the variability we see across trials may be explained by latent variables occurring in this naturally functioning system. The diverse kinematic combinations we documented mirror muscle’s adaptability and may facilitate its robust function in unpredictable conditions.

 

X-ray diffraction images were recorded as .tifs at a frame rate of 200 Hz along with simultaneous EMG (25,000 Hz). The diffraction images were annotated using the software suite Musclex and data was grouped into .csvs by trial.

The parameters for Braggs law (conversion from pixel to meter space) for this experiment are: the sample to detector distance was 2 m; the beam's wavelength was 0.1033; and the detector pixel size was 172 micrometers.

The code that we used to work with the data in the .csv files is contained in the GitHub repository 'In_vivo_Myofilament_Lattice_Kinematics'. The extensive commenting there may be helpful for working with the variable formatting of the .csvs. The documentation for Musclex also contains helpful information regarding the information contained in the annotation .csvs.

There are missing values in the .csv files. These occur when intensity peaks could not be identified during annotation. In some .csvs they are identified with an underscore prefix.