Tissue-specific gene expression plays a fundamental role in metazoan biology and is an important aspect of many complex diseases. Nevertheless, an organism-wide map of tissue-specific expression remains elusive due to difficulty in obtaining these data experimentally. Here, we leveraged existing whole-animal Caenorhabditis elegans microarray data representing diverse conditions and developmental stages to generate accurate predictions of tissue-specific gene expression and experimentally validated these predictions. These patterns of tissue-specific expression are more accurate than existing high-throughput experimental studies for nearly all tissues; they also complement existing experiments by addressing tissue-specific expression present at particular developmental stages and in small tissues. We used these predictions to address several experimentally challenging questions, including the identification of tissue-specific transcriptional motifs and the discovery of potential miRNA regulation specific to particular tissues. We also investigate the role of tissue context in gene function through tissue-specific functional interaction networks. To our knowledge, this is the first study producing high-accuracy predictions of tissue-specific expression and interactions for a metazoan organism based on whole-animal data.
| 