Regarding the excluded candidates, I want to make sure it is clear that if we excluded a candidate for some reason, dicistronic, splicing, etc., that doesn't mean we had good evidence that it was in fact dicistronic, spliced, etc. It only means that we were unable to definitively exclude that possibility. They might be readthrough, or not coding at all -- we just weren't sure enough to include in the readthrough candidates list. Also note that we were using annotations from version 5.13; we have not checked on how they have changed since then. A spreadsheet is attached that includes the transcripts rejected for the 3 reasons you mention as well as the readthrough candidates. It includes a link to the color-coded alignment, and in some cases a comment indicating which ATG we think could possibly be an IRES.