STAR★Methods

Key resources table

REAGENT or RESOURCE | SOURCE | IDENTIFIER
Antibodies

Mouse Anti-Tubulin, Acetylated | Sigma-Aldrich (Merck) | Cat#T6793; RRID: AB_477585
Rabbit anti-DsRed (Living Colors) | Takara Bio | Cat#632496
Goat anti-Mouse Alexa Fluor 647 | ThermoFisher Scientific | Cat# A-21240; RRID:AB_2535809
Goat anti-Mouse CY3 | Abcam | Cat# AB97035

Chemicals, peptides, and recombinant proteins

Coral Pro Salt | Red Sea Aquarium System | Cat# R11220
Magnesium chloride (MgCl2) hexahydrate | Sigma-Aldrich (Merck) | Cat# M2670
Magnesium chloride (MgCl2) solution | Sigma-Aldrich (Merck) | Cat# M1028
Paraformaldehyde (PFA) | Sigma-Aldrich (Merck) | Cat#158127
Dimethyl sulfoxide (DMSO) | Sigma-Aldrich (Merck) | Cat# D8418
Phospate buffered saline (PBS) | Sigma-Aldrich (Merck) | Cat# P4417
Hydrogen peroxide 30% unstabilised, AnalaR NORMAPUR® for trace analysis | WVR Chemicals | Cat# 23615.421
TritonTM X-100 | Sigma-Aldrich (Merck) | Cat# X100
Normal Goat Serum | Sigma-Aldrich (Merck) | Cat# G9023
DAPI | ThermoFisher Scientific | Cat# D1306
ProLong™ Glass Antifade Mountant | ThermoFisher Scientific | Cat# P36980

Deposited data

Nematostella vectensis 48 hpf–120 hpf body and coordinate information | This paper | https://doi.org/10.6084/m9.figshare.32820290

Experimental models: Organisms/strains

Nematostella vectensis Wildtype | Fabian Rentzsch | N/A
Nematostella vectensis elav::mOrange | Fabian Rentzsch | N/A

Software and algorithms

Arduino IDE 2.2.1 | Arduino | N/A
Fiji ImageJ | Schindelin et al.55 | N/A
TrackMate | Ershov et al.56 | N/A
MultiStackReg | Thévenaz et al.57 | N/A
FreQ | Jeong et al.58 | N/A
GraphPad Prism | GraphPad software | www.graphpad.com
R Version 4.3.1 and 4.5.0 |  | https://www.R-project.org/
Computational framework ACTIONet | Mohammadi et al.59 | GitHub - shmohammadi86/ACTIONet at R-release · GitHub
HMMER software package for sequence analysis v3.4 | Eddy60 | (http://hmmer.org/)
dplyr | Hadley Wickham et al. | https://doi.org/10.32614/CRAN.package.dplyr
ggplot2 | Hadley Wickham et al. | https://doi.org/10.32614/CRAN.package.ggplot2
ggridges | Claus O. Wilke | https://doi.org/10.32614/CRAN.package.ggridges
tidyr | Hadley Wickham et al. | https://doi.org/10.32614/CRAN.package.tidyr
purr | Hadley Wickham and Lionel Henry | https://doi.org/10.32614/CRAN.package.purrr
stringr | Hadley Wickham | https://doi.org/10.32614/CRAN.package.stringr
zoo | Achim Zeileis et al. | https://doi.org/10.32614/CRAN.package.zoo
data.table | Tyson Barrett et al. | https://doi.org/10.32614/CRAN.package.data.table
diptest | Martin Maechler | https://doi.org/10.32614/CRAN.package.diptest
broom | David Robinson et al. | https://doi.org/10.32614/CRAN.package.broom
emmeans | Russell V. Lenth et al. | https://doi.org/10.32614/CRAN.package.emmeans
Behavioral mode analysis Nematostella vectensis larval swimming (R code) | This paper | https://doi.org/10.6084/m9.figshare.32817278

Other

8-RGB LED NeoPixel Sticks | Adafruit | Cat# 1426
Mega 2560 Rev3 | Arduino | SKU# A000067
GUM ® Orthodontic Wax, unflavored | Sunstar | Cat# 723RQD
Borosilicate Glass with Filament | Sutter Instruments | Cat# BF150-86-10HP
0.2 μM PES Membrane Filtration Cup | VWR | Cat# 514-0340
Gene IDs related to Figure 4 |  | 
NV2.22582 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.24495 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.7365 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8602 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.2930 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.6626 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.7252 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.18039 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.16727 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.15467 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.10225 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.15790 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.521 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.519 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.12744 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.23620 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8726 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8930 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.1902 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.1904 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.7219 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.16659 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.7555 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.20195 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.18254 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.21369 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.20496 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.18451 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.6894 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.18932 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8942 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8357 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.18388 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.2120 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8911 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8294 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.659 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.103 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.229 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.13263 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.642 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.9107 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.8359 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.7363 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.5171 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.3562 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.7513 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.19377 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.24937 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.6623 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo
NV2.10352 | Cole et al.39 | repository [GSE200198 and GSE154105], https://www.ncbi.nlm.nih.gov/geo

Experimental model and study participant details

Nematostella vectensis wild type and Elav1::mOrange were a gift from Fabian Rentzsch (University of Bergen) and maintained after.61 In brief, sea anemones were kept in glass Pyrex dishes filled with 14 ppt artificial sea water (ASW, Coral Pro Salt, Red Sea Aquarium System, R11220) and fed 2–3 days a week with freshly hatched brine shrimp nauplii (Artemia sp.). Animals that were on a spawning cycle were kept at 18 °C in darkness. The anemones were induced to spawn regularly and after fertilization and development at RT (21 °C), larvae and polyps were used at indicated ages (48–120 h postfertilization), unless stated otherwise. Sex could not be determined for the developmental stages analyzed in this study; therefore, the influence of sex on the reported results could not be assessed. No specific ethical approval is required for Nematostella vectensis.

Method details

Behavior

Horizontal swimming setup

Swimming behavior was performed with Nematostella larvae aged 48 h postfertilization (hpf) (n larvae = 123 from N spawns = 3) 72 hpf (n = 167, N = 3), 96 hpf (n = 158, N = 3) and polyps at 120 hpf (n = 187, N = 3). In each experiment 12 larvae of a specific age were transferred into the center of a custom-made chamber (10 × 50 × 50 mm) containing 5 mL of 14 ppt ASW at RT and filmed immediately after transfer at 24 fps for 5 min using a Nikon Z50 DX 16–50 camera and a Nikkor MC 105/2.8 S lens. To illuminate the chamber evenly, four 8-RGB LED NeoPixel Sticks (Adafruit, product ID 1426) were placed at 0°, 90°, 180° and 270° at a 18 mm distance from the outer edge of the chamber (Figure S1A) and controlled through an Arduino Mega 2560 Rev3 and Arduino IDE 2.2.1 software (RGB settings: 2, 2, 2, which equals 3,0904E13 photons/cm2/s measured from the center of the arena).

Swimming and body shape changes in 2× microscopy

Animals in this experiment were 48 hpf (n larvae = 72, N spawns = 3), 72 hpf (n = 80, N = 3), 96 hpf (n = 84, N = 2) and 120 hpf (n = 80, N = 3), and experiments were performed at RT. The larvae were placed in a Low-Profile Open Diamond Bath Imaging Chamber (RC-26GLP, Warner Instruments) filled with 750 μL of 14 ppt ASW and placed on a Nikon Eclipse Ti2-U inverted microscope with a 2× objective lens. To simultaneously capture the variety of behaviors displayed and the larval body shape, two to three 15-s timelapse videos at 25 fps were obtained, using a Hamamatsu Orca camera (model C13440-20CU, light intensity 3.0. MS 2 with diffuser).

2× swimming with MgCl2

Larvae aged 96 hpf were preincubated with either 40 mM MgCl2 in 14 ppt ASW (n = 84, N = 3), or 14 ppt ASW alone (control, n = 75, N = 3) for 5 min in a 6-well plate, then transferred into the Diamond Bath Imaging chamber and immediately filmed for 15 s as described above. However, in this dataset, due to the consideration of drug exposure time, the first video was always used for analysis.

Ciliary beating frequency

All experiments in this section were performed at RT in 0.2 μM filtered 14 ppt ASW.

Wax assay

To quantify ciliary beating frequency (CBF) larvae were placed on a glass coverslip fitted with a dental wax channel (∼250 μm wide) and allowed to adjust for 2 min in ambient light followed by 30 s under microscope light before imaging. Videos of free-moving cilia of immobilized larvae were recorded with a 40× objective (Nikon Eclipse Ti2-U inverted microscope with a Hamamatsu Orca-flash 4.0 camera (model C13440-20CU) from three body regions (aboral, side, and oral) and acquired at 150 frames per second (fps), with a 3 ms (ms) exposure and 18% light intensity, maximally 90 s from the end of the adjustment period (Figures 2A and 2B). Full body images were also taken for body metric quantification.

Pipette assay

To quantify CBF over longer time periods and when exposed to 40 mM MgCl2 in 14 ppt ASW, larvae were tethered via gentle suction using a borosilicate glass micropipette pulled by a P-97 Flaming/Brown micropipette puller (Sutter Instruments). Pipettes were broken off at the tip and fire-polished to give a rounded edge and small diameter. Perfusion of filtered ASW began immediately after tethering. Images and videos were taken for CBF analysis after 5 min (CBF 1), immediately after which the perfused liquid was either maintained or switched to the 40 mM MgCl2 solution and the media in the chamber carefully replaced with the experimental liquid to ensure a consistent environment. A second set of images and videos were taken after 5 min (CBF 2) (Figures 4F and 4G). Videos were acquired at 150 fps with a 2 ms exposure and 3.2% light intensity.

Imaging

Larvae of the desired age were selected and placed individually on a glass coverslip in 2.0 μL 0.2 μM filtered 14 ppt ASW in a light squish-prep to restrict larval movement in the z-plane. Larvae were then imaged with a 20× objective (Nikon Eclipse Ti2-U inverted microscope with a Hamamatsu Orca-flash 4.0 camera (model C13440-20CU)).

Immunohistochemistry

IHC was performed after.40,62,63 Briefly, larvae were relaxed in 2.43% MgCl2 in 14 ppt ASW for 15 min, fixed in 4% cold paraformaldehyde in PBS for 1 h on a rocker with ice. Larvae were then incubated in 10% DMSO in PBS 20 min at RT, and with 2% hydrogen peroxide in PBS for 15 min at RT. Larvae were washed with PBST (Triton X-0.3%), 10–15 times, and incubated in 5% Natural Goat Serum (NGS, Merck, G9023) in PBST for 1 h at RT. Incubation in primary antibody (ABs) (Mouse anti acetylated Tubulin 1:500 (Merck, T6793) and Rabbit anti-DsRed 1:100 (Takara Bio, 632496)) in 5% NGS followed at 4 °C for 60–70 h. After subsequent washes with PBST (10–15 times), and 1 h incubation with 5% NGS, secondary ABs (all 1:200: Goat anti-Mouse Cy3 (Abcam, AB97035) and Goat anti-Rabbit Alexa Fluor 647 (Thermo Fisher scientific, A-21245)) were applied overnight on a rocker at 4 °C. Samples were subsequently washed with PBST and DAPI 1:1000 in PBS (Thermo Fisher scientific, D1306) was applied for 60 min at RT. Finally, the samples were washed 3–5 times in PBS, mounted in ProLong Glass Antifade Mountant (Thermo Fisher scientific, P36980), and imaged with a Zeiss 800 Airyscan Confocal microscope Images were processed using FIJI and Adobe Photoshop.

Single-cell transcriptomic analysis

Whole-body single-cell RNA sequencing data of Nematostella vectensis was obtained from.39 The developmental subset of the cell atlas corresponding to stages t18h to t16d was extracted for further analysis. Transcriptional pseudobulk profiles were estimated per age group using the normalized sum of total read counts over cells of each group. Read counts were normalized to counts per million (CPM) and log-transformed in base 2. Pseudobulk expression values were used to estimate temporal expression trends of individual gene markers and the average behavior of groups of genes. The cellular landscape at age td4 (approximates t96h in Figures 5 and S5; Table S25) was analyzed using the computational framework ACTIONet.59 Briefly, a low-rank approximation of the normalized count matrix is obtained using the single value decomposition (SVD). This reduced data representation is subsequently decomposed using archetypal analysis to define a low-dimensional representation for each individual cell that is useful in measuring cell similarity and building a cell manifold capturing cell relationships in transcriptomic space. The network is projected in 2D coordinates for visualization using the UMAP algorithm. All these steps were implemented using the function runACTIONet with default parameter values. To visualize gene expression levels across the cellular landscape, a network diffusion algorithm was used over the cell network to smooth genewise sparse expression values. Network diffusion is implemented in the ACTIONet’s networkDiffusion function.

Identification of putative sensory receptor ion channel genes

To systematically identify putative voltage-gated ion channel proteins, the PFAM family HMM models Ion_trans (PF00520) and Ion_trans_2 (PF07885) were used as query for searching against the reference proteome of Nematostella vectensis version NV2 (wein_nvec200_tcsv2) (https://simrbase.stowers.org/starletseaanemone). The resulting protein candidates were then annotated with the best-matching human protein (Table S31). Best human protein hits were determined by querying each putative channel sequence against the complete set of human protein-coding genes using the HMMER function phmmer. Candidates matching both Ion_trans/Ion_trans_2 families and human transient receptor potential (TRP) channels were considered as putative sensory receptor genes. All sequence searches were performed using the hmmsearch program of the HMMER software package for sequence analysis v3.4 (http://hmmer.org/).60

Quantification and statistical analysis

Statistical significance ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Different letters denote significantly different mean values, where p < 0.05.

N indicates the number of spawns and n denotes the number of animals used.

Information about statistical tests, sample exclusion, and software can be found in method details; in the related section, in the Figure legends and in the supplementary material.

Behavior

Behavioral analysis (particle tracking)

Videos from the horizontal assay (Figure 1, 5 × 5 cm arena) were analyzed using Fiji ImageJ,55 and the plugin TrackMate56,64 for particle tracking. The Hessian detector and Simple LAP tracker were employed to create particle tracks. Particle track measurements (metrics) (Figures S1B–S1I) included (I) total track distance (the sum of distances between all coordinates, where di,i+1 is the distance from one spot to the next spot in the track), (II) track mean speed (the mean of all momentary velocities within one track), (III) maximum speed (the momentary velocity with the highest value), (IV) minimum speed, (V) track displacement (the distance in a straight line between the first (di) and last coordinate (dx) of one track), (VI) maximum distance (the longest distance between any two coordinates in one track), (VIII) linearity of forward progression (a relative measurement where (VII) straight line speed is divided by (II) mean speed) and (IX) confinement ratio (a relative measurement where displacement is divided by the total distance). Data were analyzed in GraphPad Prism for Windows (GraphPad software) and R Version 4.3.1 (https://www.R-project.org/).

(I)Totaldistancetravelled=Sumi,j(di,j)

(II)Meanspeed=Mean(di,i+1×s−1)

(III)Maximumspeed=Max(di,i+1×s−1)

(IV)Minimumspeed=Min(di,i+1×s−1)

(V)Displacement=Δd=dx−di

(VI)Maxdistancetravelled=Maxi,j(di,j)

(VII)Straightlinespeed=Displacement/time

(VIII)Linearityofforwardprogression=(Straightlinespeed)/(meanspeed)

(IX)Confinementratio=Displacement/(totaldistancetravelled)

The threshold for being characterized as a stationary larva in the horizontal swimming assay was displacement <0.5 mm, and maximum distance <1.0 mm (Figures 1I and 1J). The age-specific average distance the larvae swam per minute was calculated and a smoothed conditional mean line using the geom_smooth(method = “loess”) function in R was generated (Figure 1L). See also Tables S5 and S6.

Horizontal swimming statistical analyses

Track data (metrics) and Edge data (speed information) were extracted from TrackMate. Tracks from individuals too close to another individual, where TrackMate lost the larvae and stitching was not possible were excluded from the analysis. Data were checked for normality (D’Agostino & Pearson test, and Anderson-Darling test) and found non-normally distributed for all metrics. Therefore, a Kruskal-Wallis test with Dunn’s multiple comparisons test was used as a non-parametric alternative to ANOVA (Figures S1C–S1I). To see whether the distributions of mean speed, total distance, max distance and confinement ratio (Figures 1E–1H) differed among the ages, a Kolmogorov-Smirnov test was employed. For any correlation between metrics, we used the non-parametric Spearman’s Rank correlations. See Tables S1–S4, and S7–S13 for more information.

Swimming and body shape changes in 2× microscopy quantification

By using the same particle tracking software as the horizontal (5 × 5 cm) assay, swimming tracks were generated (Figure 3D). Tracks that lasted less than 1.6 s, or from larvae that were touching each other, out of focus or partially out of frame, were excluded from further analysis. The video with the lowest number of individuals meeting the exclusion criteria was retained for analysis. To measure larval body size (Area and Axis Ratio) the “Analyze particles” tool in Fiji ImageJ was used on a median-filtered (10pixels) binary version of the videos (See Videos S1 and S2). Tracks and Edge data were extracted from TrackMate and body size measurements obtained from “Analyze particles” in FIJI. Unique larval IDs and time information from these two independent datasets were matched. Subsequently, this aligned datafile was used in combination with the behavioral state analysis again, matched, using the approximate track time. In this way, each data point was assigned a momentary speed, axis ratio, and behavioral mode.

Behavior state analysis

To categorize behavior states (Figure 3F), tracks from the aligned datafile were resampled by a threshold of 38 μm (number of coordinates before resampling = 105 324, number of coordinates after resampling = 17 789). As some stationary larvae did not generate enough datapoints, larvae with fewer than 3 datapoints in one track were not included in the following analysis (this applied to two individuals in 48 hpf (old n = 72, new n = 70)). Displacement vectors and turning angles were calculated from successive positions. Local movement (local confinement) was measured using a 20-point rolling window, and a second rolling window averaged these values (average confinement) to capture sustained behavior. Each position was assigned to one of four behavior states:

Pause: Interval between consecutive samples ≥1 s.

Swirl: Average confinement <0.50 for ≥8 consecutive positions.

Turn: Turn angle ≥10°.

Straight: States were assigned hierarchically: pause > swirl > turn > if non applied, defined as straight. Additionally, classification errors in this mode were “smoothened” by using local confinement ≥0.90.

Swimming and body shape changes in 2× statistical analyses

The relative frequency (%) of each behavioral mode was calculated per individual swimming track and a Kruskal-Wallis test with Dunn’s multiple comparisons test was employed to test whether the means varied between ages (Figure 3G, see also Table S21). Correlations between any metrics and body measurements were performed by using non-parametric Spearman’s Rank correlations (Figures 3H, S3A, and S3B, for more information see Tables S22–S24). Linear models with an Age × Axis ratio interaction were used to visualize the relationship between body axis ratio and mean speed per age and per behavioral mode (Figure 3H; Table S35).

2× swimming with MgCl2 statistical analyses

Particle tracking, body size measurements, and behavioral mode analysis were performed the same way as described above. Stationary larvae were defined as larvae with max distance <333 μm and displacement <310 μm. Differences between ratios of stationary to moving for MgCl2 data were calculated by a two-tailed two-proportion Z-test (Figure 4I). Differences in mean speed, max speed, and mean axis ratio between control and MgCl2 treated larvae were calculated by Kolmogorov-Smirnov tests and a Mann-Whitney test (Figures 4J, 4K, S4F, and S4G). For the mean speed and body axis ratio distribution data, the four quartiles (0.25, 0.50, 0.75, 1.0) were calculated in R and using these values the tracks of larvae were plotted according to which quartile they belonged to (Figures S4D and S4E). For the behavioral state analysis, one larva had fewer than 3 data points after resampling and was not included in the following analysis (this applied to MgCl2 (old n = 84, new n = 83)). Linear models using a Treatment × Axis ratio interaction combined with t-tests were used to evaluate differences in slopes between control and MgCl2 treated larvae for each behavioral mode (Figure S4I). See Tables S27–S29 and S32–S35 for more information.

Ciliary beating frequency

CBF and body metrics analysis

Analysis of recordings was conducted in FIJI ImageJ.55 Videos were aligned to correct for x-y drift using the plugin MultiStackReg57 and analyzed for ciliary beating frequency using the plugin FreQ58 (Figure 2D). Cilia length and tuft cilia length were measured from the point at which the cilium emerges from the larval body to its tip. Each point represents the average of one individual larva, comprising an average of 8–12 cilia. Length and width of larvae oriented in the horizontal plane were measured, and body axis ratio calculated by dividing length by width. Due to unequal variance between age groups, differences in body axis ratio were tested by Brown-Forsythe ANOVA test and Dunnett’s T3 multiple comparisons test (Figures 2E and 2F; Table S20). Data were analyzed in GraphPad Prism for Windows (GraphPad software, www.graphpad.com) and R Version 4.5.0 (https://www.R-project.org/).

Wax assay statistical analyses

All data were checked for normality using both the D’Agostino & Pearson test and the Anderson-Darling test. To test for differences in both cilia and body metrics over development, the following analyses were performed: Differences in CBF (48–96 hpf n = 10–18, N = 6) were assessed using a Kolmogorov-Smirnov test. Hartigan’s Dip Test was applied to assess multimodality (Figure 2D). A small number of groups within the ciliary length data were determined to be non-normal. However, due to sample size and differing variabilities, an ANOVA was determined to be the best method for statistical analysis. Therefore, for cilia length data (48–96 hpf n = 21–30, N = 5; 120 hpf n = 28–31, N = 1) within region-groups, a one-way ANOVA with Šídák’s multiple comparisons was performed; testing for differences in cilia length within age-group was also done using a mixed-effects analysis with Šídák’s multiple comparisons (results not described here) (Figure 2E). For body axis ratio (48–96 hpf n = 31–37, N = 4; 120 hpf: n = 31, N = 1), a one-way ANOVA with Dunnett’s T3 multiple comparisons was performed (Figure 3B), and, tuft (cilia) length (48–96 hpf n = 16–21, N = 5; 120 hpf n = 13, N = 1) was compared using a Kruskal-Wallis test with Dunn’s multiple comparisons (Figure 2F). For more information see Tables S15–S19.

Pipette assay statistical analyses

CBF data in pipette-attached larvae were determined to be non-parametric. To compare the control (CBF 1) and treatment (CBF 2) conditions for each treatment group, a Wilcoxon matched-pairs signed rank test was done (ASW n = 8, N = 2; MgCl2 n = 10, N = 2) (Figure 4H). See also Table S26.

Immunohistochemistry

Elav::mOrange neuronal quantification

Elav::mOrange positive neurons visualized through immunohistochemistry were counted manually for 48 hpf (n = 3), 72 hpf(n = 5), 96 hpf (n = 5) and 120 hpf (n = 6). Two independent counts were averaged and data plotted as mean ± SEM (Figure S4A). For more information see Table S30.