[
International Worm Meeting,
2017]
The development of C. elegans is precise and stereotyped, including patterns of cell division in early embryogenesis. Nevertheless, natural genetic variation in wild-type isolates can cause dramatic differences in phenotype following single-gene perturbations, indicating that different wild-type genotypes harbor functional variation in critical gene networks1. These same wild isolates also show extreme variation in the efficacy of germline RNAi1. What are the genetic, molecular and cellular mechanisms that govern these differences? And how do they evolve when stabilizing selection ensures that phenotypic development remains stable and stereotyped? Here we use single-molecule FISH to quantitatively measure the gene expression at specific locations and time points in early development. By characterizing the temporal and spatial heterogeneities of mRNA transcript numbers in the first few cell divisions, we can connect sub-cellular phenotypes to known variations in early embryonic pathway function and germline RNAi. We use a high-throughput, semi-automated pipeline to acquire precise transcript counts at precisely staged embryos, including implementation of the machine learning spot-counting software Aro2. Despite near-invariant cell division phenotypes, wild isolates show significant differences in transcript abundance for critical embryonic genes. These differences in gene expression do not fully explain differences in embryonic lethality following gene knockdown, as neither wild-type gene expression nor transcript abundance following RNAi correlates perfectly with patterns of embryonic lethality. Notably, we observe significant difference in transcript abundance variance following RNAi among wild-type isolates, suggesting inefficiency of RNAi may be controlled by stochastic thresholds. Currently, we are scaling up the experiments using a microfluidic chip specifically designed for worm embryos in order to test hypotheses with high statistical rigor. References: 1- Wild worm embryogenesis harbors ubiquitous polygenic modifier variation. A.B. Paaby, A.G. White, D.D. Riccardi, K.C. Gunsalus, F. Piano, M.V. Rockman. Elife (2015), p. 4 2- Aro: a machine learning approach to identifying single molecules and estimating classification error in fluorescence microscopy images. A.C. Wu and S.A. Rifkin. BMC Bioinformatics (2015), 16:102