This change gave rise to numerous problems and should be undone, so each CDS sequence in a coding/peptide split frame has its own distinct dataset sequence, avoiding ambiguities in residue/sequence/feature mapping and alignment.
JAL-3748CDS alignment doesn't match original CDS when aligning protein products
JAL-3700Complement sequence correspondence is broken for CDS with shared dataset sequence
JAL-3725Position calculation fails for overlapping virtual features
JAL-3613Peptide-to-CDS tracking broken for EMBL gene products
JAL-3751Overlapping CDS in ENA accession not correctly mapped by Jalview