Furthermore, 17-PA cotton homologs that are related to MIXTA, MYB5, GL2 and eight genes in the auxin, BR, GA and ethylene pathways were induced during fiber AbK initiation in the wild-type but were repressed in the naked seed mutant. Using a 2-DEbased comparative proteomics method, Liu et al. analyzed and compared the proteome profiles of -3 and 0 dpa cotton ovules between the wild-type and fuzzless-lintless mutant and found that 46 proteins showed significant differences, including many ROSrelated and stress-responsive proteins, suggesting that ROS generation increased in Xu-142-fl mutant coupled with a decreased stress response, lowered ROS scavenging capability, lowered carbohydrate metabolism, and down-regulated post-transcriptional and post-translational regulation. In combination with the transcriptomics research, proteomics analyses greatly improved our understanding of the fiber initiation defect of the Xu-142-fl mutant cotton. In this study, a precise 2-DE-based comparative proteomic analysis was performed to systematically profile the dynamic proteomes of allotetraploid cotton ovules between wild-type cotton and its fuzzless-lintless mutant during the entire process of lint fiber initiation at five distinct time points of -3, -1, 0, +1 and +3 dpa. Of the 1,800 stained protein spots that were reproducibly detected in each gel image, a total of 91 protein spots were found to be differentially accumulated in wild-type/mutant allotetraploid cotton ovules with a greater than 1.5-fold abundance dynamics between any two time points of the initiation process. The MS identification of 91 differentially accumulated proteins combined with their abundance changes as well as the results that were obtained from biochemical analyses revealed a series of different molecular events that are potentially involved in cotton fiber initiation processes. Most importantly, in contrast to previous proteomics studies, our analyses collectively revealed that the ROS burst at 0 dpa was prominent in Xu-142 wild-type cotton ovules but not in those of the fl mutant, suggesting that a high level of ROS is important for proper fiber initiation, which was confirmed through an in vivo addition experiment of H2O2 in unfertilized cotton ovules. The results not only explained the temporal-spatial co-regulation in the current functional genomics datasets but also helped to depict a full picture of the regulatory networks of essential proteins for cotton fiber initiation. In summary, the dynamic changes in abundance of these 91 DAPs may indicate that the epidermal cells of the wild-type ovules could sense the initiation signal by modulating the amount of the proteins and the corresponding molecular events, eventually transforming epidermal cells into a new shape: cotton fibers.