Among the transcribed regions of a gene, introns are the most polymorphic; therefore, they are ideally suited for developing molecular markers. The identification of intron regions, however, is not a straight-forward process and involves the alignment of expressed sequence tags (EST) or cDNA with their genomic counterpart. In cotton, this process is exacerbated by the scarcity of cotton genomic DNA sequences in Genbank. In this study, the possibility of utilizing genomic sequences from Arabidopsis, whose genome has been completely sequenced, to locate intron regions in cotton was evaluated. Cotton ESTs were searched in BLAST against the Arabidopsis database to identify orthologous genes. Cotton introns were identified with a 92% success rate, based on the alignment of cotton ESTs with Arabidopsis genomic DNA, which demonstrated that this approach is both feasible and practical for predicting the locations of introns in cotton ESTs. A majority of cotton introns had the canonical GT-AG splice site junctions, facilitating their identification in the sequence alignment process. Comparison of sequences between G. arboreum L. and G. raimondii Ulbr. indicated that introns had an almost four-fold greater variation in nucleotides than exons. A majority of the differences were due to a repeating thymine (T) or to the number of simple sequence repeat motifs.
