Quantitative developmental transcriptomes of the sea urchin Strongylocentrotus purpuratus

Abstract

Development depends on the precise control of gene expression in time and space. A critical step towards understanding the global gene regulatory networks underlying development is to obtain comprehensive information on gene expression. In this study, we measured expression profiles for the entire expressed gene set during sea urchin embryonic development. We confirmed the reliability of these profiles by comparison with NanoString measurements for a subset of genes and with literature values. The data show that ~16,500 genes have been activated by the end of embryogenesis, and for half of them the transcript abundance changes more than 10-fold during development. From this genome scale expression survey, we show that complex patterns of expression by many genes underlie embryonic development, particularly during the early stages before gastrulation. An intuitive web application for data query and visualization is presented to facilitate use of this large dataset.

Keywords: Developmental transcriptome; RNA sequencing; Sea urchin.

Figure 1

Comparison between measurements by NanoString and RNA-seq. (A) Measurements of 173 genes in nine samples of different embryonic developmental stages. (B) An example (FoxF) of time course profiles measured by NanoString and RNA-seq. The correlation coefficient between the two profiles is 0.919. Solid line: NanoString; Dashed line: RNA-seq. (C) The distribution of correlation coefficients of the time course profiles measured by NanoString and RNAseq for each gene. The median value is 0.923. See also Figure S1.

Figure 2

Various distributions of transcript abundance. (A) The highest and lowest transcript abundance for each gene during embryogenesis. The ordinate is the highest and the abscissa is the lowest transcript abundance. The ordinate cutoff (>300 transcript per embryo, horizontal dashed line) defines whether a gene has been expressed: the upper colored dots are the embryonic gene repertoire; the lower gray dots are genes which are below threshold. The abscissa cutoff (vertical dotted line) defines whether a gene has been continuously expressed: the dots on the right side represent the embryonic housekeeping genes, and the dots on the left side represent the genes whose transcript levels are very low at one or more time points. The diagonal color layers indicate the ratio between the highest and lowest transcript abundance: the blue dots represent genes expressed at constant levels with the fold less than 3; the red dots represent genes expressed at dynamic levels varying more than 10 fold. (B, C, D) Distributions of transcript abundance in the 24 hpf embryo. (B) Linear histogram of transcript abundance showing between 0–3000 transcripts per embryo. The vertical line represents 300 transcripts per embryo. (C) Log histogram of transcript abundance. All values were shifted up by 1 so that non-expressed genes (the bar at 10⁰) can be made visible on the log plot. The vertical line represents 300 transcripts per embryo. (D) Cumulative abundance.

Figure 3

Heatmap of relative expression profiles of embryonic genes. The transcript abundance of each gene is shown as a horizontal line and the lines are ordered by clusters. The ordinate shows IDs of clusters. The clusters are divided into four major groups. The abscissa gives developmental stages in terms of hours post fertilization. The density indicates the abundance relative to the maximum value for the given gene during embryonic development. Black is maximum and white is zero. See also Figure S2 and Figure S3.

Figure 4

Change of number of actively transcribed genes of each function class. Classes compiled in our earlier study (Tu et al., 2012) are annotated with the prefix ‘DB’. Classes compiled using Blast2GO in this study are annotated with the prefix ‘GO’. Highlighted (black bars) classes show an increasing trend.

Figure 5

Gene numbers for different function classes in the maternal transcriptome. Black bars represent the gene number of the given class observed; white bars represent the gene number of the given class expected according to its percentage of the whole gene set. Black triangles indicate over-representation with significant p values when comparing the observed and expected gene numbers by a binomial test. White triangles indicate under- representation. See also Figure S4.