Figure 2.

Pipeline flowchart. The pipeline proceeds through a series of steps in which data are collected and put into eight storage folders. Each storage folder feeds data into a subsequent program, which produces the next storage folder. The number of candidate introns decreases at each step, while more information accumulates for the smaller set of introns. To summarize the overall process briefly, a BLAST search identifies candidate IEPs in GenBank and DNA sequences are downloaded. RTs that are not IEPs are filtered out, and retained candidates are assigned to an intron class. ORF domains (0, 1, 2a, 2b, 3, 4, 5, 6, 7, X, En) are identified and ORF boundaries are annotated. The intron boundaries are then identified and an RNA structure is generated. Candidates with >95% similarity are grouped and a prototype from each group is identified.