GC content dependency of open reading frame prediction via stop codon frequencies

► We model the effect of varying GC content on stop codon distances. ► We describe the effect on length thresholds, error rate and E value at ORF prediction. ► Length thresholds for ORF candidates can be optimized dependent on GC content. ► We determine a maximum GC content for ORF length as a prediction feature. ► We demonstrate the applicability of our method by three example genomes.