Supplemental Figure legends.
Figure S1: Navigation diagram of ANISEED. This diagram indicates the main interfaces (on the periphery of the graph) and their links to hub pages (yellow frame) and the Genome browser (red frame). The color fillings group hubs by the type of information they display: mauve, genomic features; grey, anatomical features; beige, literature information; blue, expression data; green, gene regulatory interactions.
Figure S2: ANISEED menus. This figure shows how to access data from the different sections of the ANISEED menu. Anisearch, in the banner page and in the top right corner of each web page, is a Google-like search engine, which scans all tables of the database for the desired information. The name of the interfaces is generally self-explanatory. Note that three batch query interfaces are proposed: The genes "batch queries" interface identifies, among a list of gene IDs, those that contain specific InterPro domains, Gene Ontology terms, or are expressed in anatomical territories of interest. The anatomical "Batch Queries" identifies, among a list of anatomical territories IDs, those that have specific characteristics in terms of biometry, cell neighborhood, lineage, fate, shape and gene expression. Finally the "Query integration" interface found in the Genes section identifies ISH expression data that match both a list of gene IDs and anatomical territory IDs.
Figure S3: Anatomical ontology, fates and lineage. A) Example for each period of three anatomical ontologies at the indicated stages. Ascidians develop with an invariant lineage that has been worked out by Conklin and his successors up to the gastrula stage (Nishida 1987; Conklin 1905). Accordingly, we describe the anatomy of cleavage-stage embryos with the accepted Conklin nomenclature (left). After the onset of gastrulation, the precise lineage of each blastomere is not known and the Conklin nomenclature cannot serve as basis for the ontologies. By this stage, however, blastomeres have become fate restricted, and many of these fates are shared with other chordates. From the gastrula to the larval stages, we designed Ciona embryonic ontologies according to germ layers and fates, taking care to make them as comparable as possible to ontologies of other chordates, in particular vertebrates (e.g. Daniel Sobral et al. 2009) (middle). Following metamorphosis, the chordate body plan is lost and a tunicate-specific body organization acquired. The adult stage anatomical ontology we designed (right) was based on studies in Halocynthia roretzi (Hirano and Nishida 1997, 2000) and Ciona intestinalis (Tokuoka et al. 2005). Green lines link territories belonging to the same lineage. B) Recent published refinements of anatomical information. The B7.6 cell pair in Ciona intestinalis, initially thought to divide first after metamorphosis and contribute germ cells, was recenty shown to undergo an asymmetric cell division during gastrulation to produce B8.11 (endodermal strand) and B8.12 (germ cells) (Shirae-Kurabayashi et al. 2006). In addition, a recent analysis showed that Ciona intestinalis B7.5 cells give rise to trunk ventral cells (TVC) and muscle (Satou et al. 2004), but do not contribute to endoderm, as initially proposed in Halocynthia (Nishida 1987). Finally, b8.19 contributes tail-tip muscle cells in Halocynthia roretzi but not in Ciona intestinalis (Nishida 1987). These deviations from the Conklin lineage are relatively minor, but they suggest that the ascidian lineage may slightly vary between species, and that anatomical ontologies for each species will evolve as technology progresses. C) Refinement of anatomical information through the analysis of 3D embryo. The figure shows reconstructed Ciona intestinalis embryos just before invagination. A7.6 has already cleaved and thus counts 112 cells, rather than 110 in Styela (Conklin 1905) and Halocynthia (Nishida 1987). Left, antero vegetal view of an initial 112-cell stage embryo. Solid colors indicate the fates at the larval stage of vegetal blastomeres: blue, endoderm; orange, notochord; yellow, nerve cord; green, trunk lateral cells; red, tail muscle and mesenchyme. Right, the same embryo and angle of view, in which the A7.6 daughter cells only are colored.
Figure S4: Examples of 3D embryo reconstructions. Reconstructed Ciona intestinalis embryo models from 64-cell to gastrula stages. Animal cells are represented as a mesh. Top panels show a vegetal view, bottom a vegeto-lateral one. Anterior is to the left. Code color is the same as in Fig S3.
Figure S5: The NISEED functional annotation pipeline. Data stored in NISEED are represented in grey. These data are generated by the integration in the pipeline of external software represented in Orange. This annotation procedure associated to individual ANISEED gene models Interpro IDs (75% of genes annotated) and GO terms (60%, 51% and 45% of genes received at least one GO term describing their molecular function, biological process or cellular component, respectively). Gene names and GO annotations inferred from evolutionary conservation can be manually edited and complemented using NISEED-Curator tools on the basis of experimental evidence gained in the model organism studied.
Figure S6: Classification of cis-regulatory sequences. The figure indicates the definitions used in ANISEED for different types of cis-regulatory sequences and corresponding terms in sequence ontology (Eilbeck et al. 2005). The schemes illustrate examples of each class of sequence.
Figure S7: Unequal cleavages between fertilization and the onset of gastrulation. The mother cell of each unequal division is colored according to the legend. For each unequal division, a thicker line links the mother cell to its larger daughter. Asymmetric divisions shown were found in all Ciona intestinalis embryos analyzed, though the amplitude of the asymmetry could vary between embryos.
Figure S8: In silico identification of the anterior neural inducer. Two sets of independent molecular (blue box) and anatomical (green box) queries are submitted, leading in both cases to the identification of sets of molecular or anatomical IDs, which are combined (red box queries) into a search for genes from the molecular query expressed in the territories identified by the anatomical query. The virtual embryo pictures displays the expression pattern at the 32-cell stage of FGF9/16/20 (pink), and the position of the a6.5 blastomeres (green).
Figure S9: Automatic inference of regulatory interactions significantly extends the Ciona developmental Gene Regulatory Network. This network (total interactions represented as blue bars at each stage) includes the network from Imai et al. (2006)(red bars), and extended it significantly with regulatory interactions extracted from published small-scale studies.