The development of bioinformatic solutions is guided by biological knowledge of the subject. In some cases, we use unambiguous biological models, while in others we rely on assumptions. A commonly used assumption for genomes is that related species have similar genome sequences. This is even more obvious in the case of chloroplast genomes due to their slow evolution. We investigated whether the lengths of complete chloroplast sequences are closely related to the taxonomic proximity of the species. The study was performed using all available RefSeq sequences from the asterid and rosid clades. In general, chloroplast length distributions are narrow at both the family and genus levels. In addition, clear biological explanations have already been reported for families and genera that exhibit particularly wide distributions. The main factors responsible for the length variations are parasitic life forms, IR loss, IR expansions and contractions, and polyphyly. However, the presence of outliers in the distribution at the genus level is a strong indication of possible inaccuracies in sequence assembly.
Keywords: chloroplast genome; genome databases; sequence length; taxonomy.