Digital gene expression for non-model organisms

We present a method for digital gene expression profiling called EDGE, or EcoP15I-tagged Digital Gene Expression, based on ultra high-throughput sequencing of 27 bp cDNA tags using the Illumina Genome Analyzer platform. Since there is a one-to-one correspondence between tag and gene in EDGE, it is particularly suited for quantification of transcript abundance in non-model organisms where a high quality annotated genome is not available. We show that EDGE is capable of assaying for expression in >99% of genes in the genome and achieves saturation in detecting transcripts after 4 – 8 million sequence reads. In addition, EDGE exhibits little noise from library construction and the flow cell sequencing process, and reveals a large (106) dynamic range of gene expression. In a direct comparison with RNA-Seq, we demonstrate that both methods provide similar assessments of relative transcript abundance. However, unlike RNA-Seq, EDGE is robust in detecting gene expression differences for poorly expressed genes and does not exhibit transcript length bias. Leveraging the increased sensitivity of EDGE compared to microarrays, we applied it to a laboratory mouse model for pigment-type switching and found that the skin and spleen of animals carrying the defective melanocortin-1-receptor displayed a downregulation of genes involved in the interferon response, suggesting a previously unappreciated role for the Mc1r in innate immunity. Finally, we illustrate various strategies for making tag to gene assignments in a non-model organism, the cheetah, and provide new biologic insight into mammalian pigment patterning.