A digital framework to build, visualize and analyze a gene expression atlas with cellular resolution in zebrafish early embryogenesis

Abstract

A gene expression atlas is an essential resource to quantify and understand the multiscale processes of embryogenesis in time and space. The automated reconstruction of a prototypic 4D atlas for vertebrate early embryos, using multicolor fluorescence in situ hybridization with nuclear counterstain, requires dedicated computational strategies. To this goal, we designed an original methodological framework implemented in a software tool called Match-IT. With only minimal human supervision, our system is able to gather gene expression patterns observed in different analyzed embryos with phenotypic variability and map them onto a series of common 3D templates over time, creating a 4D atlas. This framework was used to construct an atlas composed of 6 gene expression templates from a cohort of zebrafish early embryos spanning 6 developmental stages from 4 to 6.3 hpf (hours post fertilization). They included 53 specimens, 181,415 detected cell nuclei and the segmentation of 98 gene expression patterns observed in 3D for 9 different genes. In addition, an interactive visualization software, Atlas-IT, was developed to inspect, supervise and analyze the atlas. Match-IT and Atlas-IT, including user manuals, representative datasets and video tutorials, are publicly and freely available online. We also propose computational methods and tools for the quantitative assessment of the gene expression templates at the cellular scale, with the identification, visualization and analysis of coexpression patterns, synexpression groups and their dynamics through developmental stages.

Figure 1. Schematic description of the atlas construction process.

For each developmental stage, the partial 3D volumes of the analyzed embryos and the 3D volume of the whole template embryo were processed for nuclear center detection and gene pattern segmentation. Mapping the analyzed embryos onto the corresponding common template was guided by the specimen's shape, revealed by the nuclei, and by the segmented gsc expression pattern, chosen to be a common reference. Each step was supervised and, if necessary, corrected via an interactive graphical user interface. The final model, where all the gene patterns coming from different individuals could be jointly compared, constitutes one 3D atlas template. The Match-IT software performs the gene pattern segmentation and the validated mapping. The Atlas-IT software allows interactive visualization of the 3D atlas template.

Figure 2. 3D raw data, nuclear center detection, gene pattern segmentation and their validation at 6.3 hpf.

(a) Upper panel: volume rendering, lower panel: axial orthoslice of an analyzed embryo's nuclei (white), reference gsc pattern (red) and ntla pattern (green). (b) Same with template nuclei (blue) and their gsc pattern (red). (c) Nuclear positions (yellow) superimposed on the raw nucleus images (white) displayed by three orthoslices in the , and planes. (d) Zoom on the template gsc raw expression (red) superimposed on the template nuclear positions (blue). (e) 3× zoom on the boxed region in (c) with detected nuclei positions (pale yellow), an example of a validated nucleus (green), a false positive (red), a false negative (yellow) and a selected position to be evaluated (white cube). (f) Same as (d) with the segmented gsc domain (white). Scale bars, m. Axes point to the animal pole (), dorsal side () and lateral side () of the embryo respectively.

Figure 3. Mapping procedure in the 6.3 hpf atlas template.

(a) Analyzed embryos: detected nuclei (white), gsc positive cells (red), automated initialization scheme extracting the plane passing through the blastoderm margin (green), bilateral symmetry plane (purple), and referential . (b) Same with the template, detected nuclei in blue. (c) Initialization step aligning the basis of the analyzed embryo and the template; the yellow arrowhead points to a mismatch refined in (d) through the registration procedure. Scale bar m.

Figure 4. Exploring the 3D atlas with the visualization tool Atlas-IT.

(a) Atlas-IT interface displaying the template nuclei (light blue), segmented gene expression patterns of ntla (blue) and flh (green). (b) From left to right: equatorial, sagittal and dorsal views of the 9 individual gsc boundaries as compared to the mean gsc domain (red) at 6.3 hpf. (c) Evolution of the oep-gsc pair over time after being mapped onto the template. (d) Evolution of the oep-gsc pair over time in the analyzed embryo where they were co-stained. Scale bar m.

Figure 5. Assessing gene coexpression and cell genetic profiles.

(a) Matrix displaying the percentage of cells coexpressing any given gene pair at developmental stages from 4 to 6.3 hpf (see also Fig. S21). Gene pairs were ordered according to the similarity of the evolution of their patterns over time (Fig. S23). (b) Spatial WPGMA clustering at 6.3 hpf of the 1,194 positive cells according to the similarity of their gene expression profiles. (c) Volume rendering, (d) lateral view and (e) coronal view of template nuclei at 6.3 hpf. Nuclei were classified according to their gene expression profiles, which revealed 5 distinct morphogenetic domains: dorsal hypoblast (yellow), marginal dorsal epiblast (blue), dorsal epiblast (white), paraxial and lateral blastoderm margin (red), forerunners and dorsal YSL (green). White arrowheads indicate the limits of the imaged analyzed embryos. Upper panel, sagittal section, lower panel equatorial section passing through the embryonic shield. Scale bar m.