The loss of previously established genes has been proposed as a major force in evolutionary change. While genome sequencing of many new species offers the opportunity to identify cases of gene loss, it is unclear which algorithms offer the greatest accuracy or sensitivity. A number of methods to identify gene losses rely on the presence of a pseudogene for each loss. If genes are deleted when lost, however, such methods will fail to identify these cases. As the fate of gene losses is still unclear, we identified gene losses through a method that does not require pseudogenes to identify human-specific gene losses. Of the several hundred probable gene losses initially identified, we were unable to find a single case of unambiguous gene loss via deletion. We were also able to identify a large number of previously unannotated genes in the human genome, some of which also had evidence for transcription. Though our results suggest that pseudogene-based methods for finding gene losses in humans will not miss many events, we discuss the dependence of these conclusions on the divergence times among the species considered. Supplementary Material is provided (see online Supplementary Material at www.liebertonline.com ).