Fig 1.
Adult C. elegans swimming is low dimensional.
(a) Schematic showing the conversion of a worm body posture into curvature angles at a single time point. The shape of the worm is divided into 11 segments along the anterior-posterior axis and angles are calculated between adjacent segments. The 10 angles are converted to a colormap with blue-to-red indicating dorsal-to-ventral deflections. (b) Kymograph of body curvature of a day-1 adult worm during 10-seconds of swimming. Body segment number is plotted on the y-axis and time on the x-axis. (c) The variance in swimming postures of adult worms is largely captured by a linear combination of four eigenworms. The first four eigenworm shapes are reconstructed and are outlined in blue, orange, green, and red, respectively. (d) The fraction of the total variance explained in swimming postures as a function of the number of eigenworms used in reconstruction of adult C. elegans (n = 43) swimming behavior. Dashed colored lines indicate the cumulative variance explained by the first four eigenworms at 53%, 85%, 94%, and 97%, respectively.
Fig 2.
C. elegans adult swimming and crawling are distinct gaits.
(a) Schematics of the first two principal eigenworms from adult swimming as shown in Fig 1C. (b) Swimming eigenworm amplitude distributions show a stereotyped ring structure which captures the coordination of swimming eigenworms one and two to produce swimming locomotion in adult C. elegans (n = 43). (c) Schematics of the first two principal eigenworms of adult crawling behavior. (d) Crawling eigenworm amplitude distributions show a stereotyped ring structure of coordination between crawling eigenworms one and two in adult C. elegans (n = 38). (e and f) Polar plots of eigenworm amplitude (b and d) speeds as a function of phase in the ring. Speed data across all animals are plotted as scatter points, and the mean is overlayed. (e) In swimming, speed is bimodal and is slowest when in the “C” shape, or eigenworm one, whereas in crawling (f) the speed is constant along the ring. (g) 3D scatter plot of the first three eigenworm amplitudes from a representative worm tracked during a swimming-to-crawling transition. The ring structure of swimming (blue) is distinct from the ring structure associated with crawling (orange).
Fig 3.
Rhythmic swimming is present at birth and matures throughout development.
(a) Swimming eigenworms 1–4 across developmental stages: young L1 (blue), late L1 (orange), L2 (green), L3 (red), L4 (purple), adult (brown). (b) Participation ratios (PRs) representing the dimensionality for each swimming tracking session of young L1 (n = 86), late L1 (n = 40), L2 (n = 47), L3 (n = 48), L4 (n = 39), and adult (n = 43) C. elegans. Young L1 and adult C. elegans swimming PRs show a significant difference in means (p = 9.07e-08, t-test). (c-g) Swimming locomotion represented by eigenworm one and two amplitude distributions across developmental stages: young L1 (c), late L1 (d), L2 (e), L3 (f), and L4 (g) demonstrate coordination of these eigenworms is present across development, however young L1 worms also produce uncoordinated postures not represented by the first two eigenworms. (h) The developmental stages, young L1, late L1, L2, L3, L4 of N2 C. elegans recorded in this study. Dashed lines in (b) represent means and interquartile range. Statistical significance in (b) was determined using Bonferroni adjusted alpha levels of 0.03 (0.05/15). Young L1 PRs showed ****p statistical significance compared to all other stages. Significance: *p<0.0033, **p<0.00067, ****p<0.0000067.
Fig 4.
C. elegans crawling postures are disrupted at birth and stabilize by late larval 1 stage.
(a) Crawling eigenworms 1–4 are identical across developmental stages: late L1 (orange), L2 (green), L3 (red), L4 (purple), adult (brown), except for posterior deviations along the body at the young L1 stage (blue). (b) Participation ratios (PRs) for each crawling tracking session of young L1 (n = 61), late L1 (n = 51), L2 (n = 51), L3 (n = 43), L4 (n = 47), and adult (n = 38) C. elegans. Young L1 and adult C. elegans crawling PRs show a significant difference in means (p = 1.87e-14, Welch’s t-test). (c-g) Crawling eigenworm one and two amplitude distributions across developmental stages: young L1 (c), late L1 (d), L2 (e), L3 (f), and L4 (g) demonstrate the coordination of these eigenworms is not present at the young L1 stage but develops by the late L1 stage. Dashed lines in (b) represent means and interquartile range. Statistical significance in (b) was determined using Bonferroni adjusted alpha levels of 0.03 (0.05/15). Young L1 PRs showed ****p statistical significance compared to all other stages. Significance: **p<0.00067, ****p<0.0000067.