Transcription changes dramatically during embryonic and larval development, as evidenced in whole embryo RNA-seq data (Gerstein et al, 2014). Likewise, mixed age populations of cell type specific cells using SAGE (Meissner et al, 2009), and microarray (Spencer et al, 2011), show differences in gene expression. We are producing a more comprehensive dataset, providing gene expression information for individual cell types from time synchronized populations of those cells throughout embryonic development.As part of the Expression Patterns in Caenorhabditis (EPiC) project (Murray et al, 2012), we have been producing fluorescent reporter strains for transcription factors (TFs) in the worm, and utilizing 4-D microscopy and cell lineaging to obtain real time expression levels for a single gene, in each embryonic cell
(http://epic2.gs.washington.edu/Epic2/). Using these strains and FACS, we are isolating time synchronized, cell type specific cells at regular intervals throughout the first half of embryonic development. Choosing reporters with the earliest possible expression allows for isolation of cells at all stages of embryonic development, and our movies identify exactly which cells we are collecting at each time point. Using RNA-seq, we have determined the mRNA content of the ABa sublineage (
tbx-37), and muscle (
hlh-1,
hnd-1), intestinal (
end-1), pharyngeal (
pha-4, excluding
end-1), and hypodermal cells (
elt-1) as they develop over time, with the goal of covering all of the major lineages and cell types in the embryo. Dramatic differences in gene expression occur both between cell types, and over time within the same cell population. Starting with the body wall muscle lineages, we intend to use our RNA-seq data to determine the order and specificity of expression of all TFs, and pair this with ATAC-seq information to infer regulatory networks (Neph, 2012). We are also utilizing mutations and RNAi to look at the downstream transcriptional changes in muscle cells after loss of key TFs. We see hundreds of genes with changes in transcription after loss of
hlh-1, including both known muscle genes, and genes with no previous association with
hlh-1 or muscle cells. Combined with HLH-1 binding site data from ChIP-seq experiments (Araya et al, 2014), we see decreased expression of genes that have binding sites for HLH-1 as well as some genes that do not have detected binding sites. This suggests both direct and indirect effects of the mutation. By combining these various data sets, we expect to construct a robust model of the TF regulatory cascade in the major cell lineages of the worm.