Comparison and development of EST-SSRs from two 454 sequencing libraries of Gossypium barbadense

EST–SSRs of Gossypium barbadense are mainly developed using traditional Sanger sequencing. However, due to the high cost and low throughput of Sanger sequencing, it is necessary to use high throughput sequencing technology for the development of more ESTs to more effectively analyze the structure and function of this species. In this study, a G. barbadense acc. 3–79 unnormalized fiber cDNA library (219.63 Mb) and a G. barbadense cv. Hai7124 normalized root cDNA library (204.61 Mb) were obtained by 454 sequencing. EST–SSRs were identified from the two libraries, and only 7,255 SSRs were obtained from the unnormalized library, with an average frequency of 1/31.00 kb. In contrast, 16,087 SSRs were obtained from the normalized library, with an average frequency of 1/13.02 kb. The frequencies of dinucleotides and tetranucleotides in the two libraries were very different. Comparing the two libraries, we found that a normalized cDNA library is more efficient for mining SSRs. Integrating the two libraries allowed the development of 1,129 EST–SSR markers, and 311 polymorphic loci were integrated into our interspecific BC1 genetic linkage map. The mapping results showed that the distribution of EST–SSRs on sub-genomes and chromosomes was uneven; however, the distribution of the mapped G. barbadense EST–SSRs on homologous chromosomes was similar, with the exception of Chr05 versus Chr19 and Chr12 versus Chr26. This study provided new EST–SSR markers that will facilitate studies on cotton genetics and breeding.