This talk is to present a methodology to estimate genome distances. We formulate the genome model of gene family evolution, under the framework of birth-death stochastic process, and then define a new additive genome distance. Because of the insufficiency of the gene-content data, we develop a novel approach to estimating the genome distance, based on the extended gene content, i.e., in a genome the status of a gene family could be either absence, presence as single-copy, or presence as duplicates. Simulations show that the new method for tree-making is efficient, consistent, and fairly robust. We apply it to 13 microbial genomes. The result supports the concept of universal tree of life.