Mike the Mad Biologist writes on his blog that there's some not so high-quality genome annotation in GenBank. Annotation, he says, hadn’t been the rate-limiting step until recently in sequencing and now that it is, many labs are turning to automated gene prediction software. However, he adds that software correctly calls about 95 percent of genes and that not all labs have check in place to find bad calls. “This is bad for people who want to pull these genomes out of GenBank and study them,” Mike writes. “But this also screws up gene identification of new genomes. High quality gene prediction processes rely on both ab initio rules and previously identified genes from other genomes.”