Description
Hair follicles are self-renewing organs within the skin which cycle through periods of growth and destruction, with an intervening period of outward quiescence. The hair follicle cycle is driven by Hedgehog and Wnt signaling and affects epithelial thickness, melanin production, immune function, and tumor susceptibility. We have previously shown that somatic alterations to the genome affect the genetic architecture of the skin. This study examines how the hair follicle cycle affects gene the genetic architecture in vivo by genomic and genetic analysis of 343 genetically heterogeneous mice during the hair follicle growth phase (anagen) and quiescent phase (telogen). We use eQTL analysis and differential correlation to identify changes in metabolic and stem cell activity not detected by differential expression. Germline influence in gene expression is profoundly higher during anagen, but this increase is not a simple factor of higher levels of gene expression. The most strongly induced eQTLs were involved in cellular energy metabolism and melanogenesis rather than hair follicle growth or hedgehog signaling. We demonstrate that hair follicle and circadian rhythm pathways are sexually dimorphic, but do not find evidence for an effect of sex on eQTL networks. We also use eQTL gene network analysis to identify candidate causal relationships between expression of genes in the hair follicle and melanin pathways, identifying Mcoln3 as a candidate for the familial melanoma locus on 1p22. To lower the bioinformatic barriers to eQTL network analysis we produced CARMEN, a free open-source stand-alone software package. This study demonstrates how to perform a systems genetic analysis of a heterogeneous tissue studied in vivo under physiologically relevant growth signals.