CD133 is regarded as a marker of tumor initiating cells in many tumors, including colorectal cancer. O'Brien and Ricci et al. have proved that in primary colorectal tumors there are colorectal tumor stem cells (initiating cells) which are marked by CD133 antigen. Using a genetic knockin lacZ reporter mouse model, Shmelkov et al. challenged this increasingly influential viewpoint and drew two important conclusions that challenge former opinions. First, CD133 is widely distributed throughout the full range of tumor epithelial cells in the colon as opposed to being limited to a few cells. Second, CD133 negative cells of colon tumors are also tumorigenic, and are more inclined to metastasize. Based on these two opinions, we hypothesize that the expression of CD133 is different among tumor cells, and that quantitative but not qualitative analyses of CD133 abundance are necessary to determine the relationship between CD133 expression and tumor stem cell characteristics. To verify this hypothesis, colorectal cancer cell line SW620 was cultured and sorted into CD133^Hi, CD133^Mid and CD133^Low subgroups using magnetic microbeads to compare their xenograft biological characteristics. The results showed that the CD133^Hi subgroup of SW620 is more close to the tumor initiating cells in terms of biological characteristics than CD133^Mid and CD133^low subgroups, but the CD133^low subgroup still maintains the ability of tumorigenicity. It supported that tumor initiating cells are more correlated to the abundance of CD133.