Single-copy orthologs are often used to reconstruct phylogenetic trees of life. One can prepare a set of single-copy orthologs by building a new database including the species/strains of interest and performing a homology search, but these steps are time-consuming if working with a large number of samples. To solve this problem, we need more efficient and practical approaches. Here we developed a new pipeline called CUSCO, which reconstructs a phylogenetic tree from the genome sequences of desired samples plus a reference set of protein sequences. The pipeline is implemented with a function to identify a minimal set of genes to reconstruct a species tree, which is comparable to the one reconstructed from single-copy orthologs. Compared to the existing software, the CUSCO pipeline finishes in a shorter time in reconstructing a species tree.