(A) An image of a threefold embryo in the twisted state, showing the untwisting markers. (B) The same image as in (A) with the untwisting markers labeled. Asterisks mark seam cell nuclei, and the …
(A) Evidence for helical twisting, highlighted on four pairs of consecutive seam cell nuclei. If no helical twisting occurs, yellow lines (connecting seam cell nucleus pairs) should appear parallel …
Coarse features such as seam cell nuclei are visible in single view iSPIM (A), but finer features such as junctions between hypodermal cells labeled with DLG-1::GFP are better resolved in the diSPIM …
(A) XZ and YZ views of an untwisted worm embryo using a lattice comprised of every other seam cell nucleus, a total of 12 points. This lattice fails to capture bends in the animal and does not …
(A) The twisted L2 larval volume displayed in the MIPAV volume renderer. (B) The twisted L2 larva after lattice-building. (C) The L2 larval worm after untwisting. See also Video 7. MIPAV, Medical …
Distances between seam cell nuclei (left) and pharyngeal lengths (right) were compared in twisted (A) and untwisted (B) worm embryos. All scalebars: 10 µm. (C) Comparative 3D distance measurements …
Comparative 3D distance measurements of seam cell nuclei pairs H0 and T (left graphs) and pharyngeal lengths (right graphs) for six embryos. In all cases, distance measurements in the twisted case …
Selected volumetric timepoints pre (A–D) and post (E–H) untwisting, with canonical state of embryo indicated at bottom. See also Video 2. (I) Cartoon of untwisted embryo, indicating coordinate …
Comparative timepoints were selected based on the H1R seam cell shifts. Max projections of volumetric images are shown. Note the underlying similarity in overall shape across animals. Scalebar: 5 μm.
Comparative timepoints were selected based on the H1R seam cell shifts. Max projections of volumetric images are shown. Note the underlying similarity in overall shape and seam cell positions across …
Before fitting, raw data are treated to remove obvious outliers (top row) and to fill in missing data (mid, bottom rows). In both cases, outliers and ‘gaps’ within data are found manually, and …
(A,B) Axial seam cell nuclear trajectories from different embryos are similar in shape, but shifted in time. (C,D) Shifting in time aligns the trajectories. (E, F) Averaging the shifted …
Data from two embryos are shown before (top) and after (bottom) temporal alignment. The data derived from embryo 4 was shifted 5 timepoints to the right, following the procedure described in 'Materia…
Different fitting models (see also Table 2) for embryonic axial displacement are plotted (red curves), against raw data (blue diamonds). Also shown on each plot are quantitative measures of goodness …
Comparisons in axial position vs. time for a seam cell nucleus (right H1, upper graph) and for CANL (lower graph). For most nuclei, as in the upper graph, positions were stereotyped to within 4.6 μm …
Examples of raw, averaged data (derived from 4 to 5 embryos, blue dots) and fits (black lines). Linear, power, and three-parameter logistic curve examples were taken from the right H0 seam cell …
(A) Snapshots of the elongating embryo near start (Volume 30, left) and end (Volume 113) of elongation. Seam cell nuclei volumes are indicated as filled spheres, L/R axes are as indicated, seam cell …
(A–D) Snapshots of the elongating embryo at start (above dashed line) and end (below dashed line) of elongation, as shown in lateral (X motion, A) and dorsal-ventral (Y motion, C) views. Distances …
(A) Early (left) and late (snapshots) in the elongating embryo. Gray spheres: seam cell nuclei; ALA cell body: blue sphere; ALA neurites: blue lines; AIY cell bodies: yellow spheres; CAN cell …
(A) Axial distance (measured from the origin point) of the ALA cell body for two ALA datasets. Similar to seam cells, axial distance increases during elongation and then plateaus once elongation has …
(A) Exemplary data for a twofold embryo. Left column: raw data. Right column: segmented data. The red neuron is RMED, the orange neuron is ALA, and the purple neurons are the cell bodies of the AIY …
The markers are yellow, indicating that fewer than 20 seam cell nuclei have been labeled.
The user shifts seam cell nucleus #9 over, adds markers for both of the 10th seam cell nuclei, shifts nucleus #6 over and adds nucleus #20. There are now 20 seam cell nuclei marked, as indicated by …
The user has increased the opacity of the volume to better enhance the appearance of the nose, now labeled in yellow.
A potential pair is found, with the mid-point marked in red. A third seam cell nucleus is found closer to the mid-point than the pair, invalidating the pair.
The correct lattice is listed first as it had the highest rank.
The user rotates the volume to get a better view during this phase. The magenta, red, and green curves represent the left- hand curve, center-line curve, and right-hand curves respectively. The …
Each of the 10 seam-cell pairs is marked, and 8 additional pairs have been added to capture the curve of the worm.
The ellipses fit within the boundaries of the natural spline curves.
The original ellipses are expanded until they contact an adjacent surface of the worm or they reach the boundary of the sample plane.
The outlines in the bottom three panels show how the surface encapsulates the volume data.
Each pair has a unique color value, indicating which pairs belong on the same slice in the final straightened image. Labeling the lattice pairs this way helps disambiguate voxels with potential …
This animation provides a graphical representation of the computational steps used to segment seam cells, build a lattice, and straighten embryo volumes. For additional information refer to Supplemen…
Despite errors in individual untwisted volumes, the overall pattern of embryonic development and elongation is clear.
The positions shown in the rendering are averaged, fitted values derived from five embryos, using the averaging and fitting procedure described in the text; the rendering thus represents a …
As in these videos, all positions are averaged, fitted values derived from multiple embryos. View is from dorsal perspective. Red spheres represent CAN cell bodies, yellow spheres represent AIY cell …
The image was imported into ImageJ and the Magenta LUT was applied to the stack. The volume shown here corresponds to the untwisted volume in Figure 1—figure supplement 4.
The volume was segmented and rendered in Imaris.
Fitting functions tested for describing axial displacement. Equations are used in Figure 4—figure supplement 2. L: length; t: time. Other parameters and their meaning are listed in the table. For …
Fitting type | Equation | Parameters |
---|---|---|
von Bertalanffy | L = A(1-exp[-B(t-C)]) | A: upper asymptotic length B: growth rate C: time at which L = 0 |
Exponential | L = A-(A-B)exp(-Ct) | A: upper asymptotic length B: lower asymptotic length C: growth rate |
Three-parameter Gompertz | L = A[exp(-exp(-B(t-C)))] | A: upper asymptotic length B: growth rate C: time at which L = 0 |
Three-parameter logistic | L = A/[1+exp(-B(t-C))] | A: upper asymptotic length B: growth rate C: inflection point |
Four-parameter Morgan Mercer Flodin | L = A – (A-B)/(1+(Ct)D) | A: upper asymptotic length B: length at t = 0 C: growth rate D: inflection parameter |
Four-parameter logistic | L = B + (A-B)/{1+exp[(C-t)/D]} | A: upper asymptotic length B: lower asymptotic length C: growth rate D: steepness parameter |
Fitting functions for each cell type. X, Y, Z trajectories were fitted as indicated functions of time (t).’ 50-point smoothing’ refers to smoothing the input data with a 50-point span, using …
Cell type | X fit | Y fit | Z fit |
---|---|---|---|
Seam cell nucleus | Power X = atb+c | Linear Y = p1*t + p2 | Three-parameter logistic Z = A/(1+exp(-B(t-C))) |
CANR/L | 50-point smoothing | 50-point smoothing | Three-parameter logistic Z = A/(1+exp(-B(t-C))) |
AIYR/L | 4th degree polynomial X = p4*t4+p3*t3+p2*t2+p1*t+p0 | Linear Y = p1*t + p2 | Three-parameter logistic Z = A/(1+exp(-B(t-C))) |
ALA ALA xR1/xL1 ALA xR2/xL2 | Linear X = p1*t + p2 | Linear Y = p1*t + p2 | Three-parameter logistic Z = A/(1+exp(-B(t-C))) |
Tutorial for use of the WormUntwisting automated lattice-building plugin.
Deviations between embryo datasets.
For each cell studied in this paper, data from 5 embryos were shifted as discussed in the text, and the standard deviations between embryo positions at each timepoint computed. Mean standard deviations (<σ>) and the maximum standard deviation (Max(σ)) over all timepoints are recorded above. For x and y coordinates, the majority of embryo positions are within 2 μm. For z coordinates, most embryo positions lie within 10 μm of each other, with the exception of CANL. See also Figure 4—figure supplement 3 for representative embryo trajectories. For most data displayed here, at least three embryo datasets were used in generating these values. For three datasets (red italics), only two embryo datasets were compared.
Deviations between fits and averaged data.
For each cell studied in this paper, the absolute differences between averaged coordinates and fits were computed at each time point. The means and standard deviations of these differences over time, in μm, are recorded in the table above. For x and y coordinates, the majority of fitted points lie within 1.5 μm of the averaged data, regardless of cell type. For z coordinates, the majority of fitted points lie within 7.5 μm of the averaged data, with the exception of CANL.
Raw annotation data for seam cell nuclei, neuronal cell bodies, and ALA neurites.
Supplementary data file 4 contains raw annotation data generated by the untwisting algorithm for the 20 seam cell nuclei; CAN, AIY, and ALA cell bodies; and ALA neurites. Each sheet contains positional information for one cell, broken up by embryo dataset. Embryo datasets are labeled in the form Embryo_#_X_minutes, where # corresponds to the number assigned to the dataset (1–8) and X represents the imaging frequency (between volumes) in minutes. For each embryo dataset, the volume numbers and X, Y, and Z-positions of the cell or neurite in that volume are listed.
The data are provided in raw form, after sorting by embryo, cell, and volume but before cleaning, shifting, and fitting. For some volumes annotation information was not captured, usually due to errors in the untwisting process; for these volumes the spreadsheet entries are left blank. Additionally, there is unconstrained rotation around the midline in most datasets, which can cause X and Y-values to switch between positive and negative sign. The canonical orientation of the embryo for this paper is for cells on the right side (R) of the animal to have positive X-values and Y-positions located dorsal to the midline to have positive values; in volumes where the XR values are negative the sign should be changed, as well as the corresponding sign for the YR, XL, and YL values for that volume. Z-measurements are insensitive to this rotation. All annotations are in μm.
Quality control measurements.
The data provided in this supplementary data file correspond to the quality control measurements used to generate Figure 2 and Figure 2—figure supplement 1. The data are sorted by embryo, volume, and measurement type. Embryos are named in the form Embryo_#_X_minutes, where # corresponds to the number assigned to the dataset (1–8) and X represents the imaging frequency (between volumes) in the dataset. All data are listed in μm.