In the recent years,massively parallel cDNA sequencing(RNA-seq)has become the most widely used technology for transcriptome profiling due to its single-base level accuracy and high-throughput speciality for detecting larger number of transcripts.In this study,we developed a computational approach and applied it on an integrated RNA-seq data set across 15 normal mouse tissues,and consequently assigned 8,408 House-keeping genes(HK genes)and 2,581 Tissue-specific genes(TS genes)among the total 23,374 annotation genes from Refseq on the basis of their expression data in distinct tissues,respectively.By comparing some basic genome features between our defined HK genes and TS genes,we found that HK genes are generally longer,have more exons,fewer repeated elements and more evolutionary conserved than TS genes.