Towards Automatic Detecting of Overlapping Genes - Clustered BLAST Analysis of Viral Genomes

Overlapping genes (encoded on t he same DNA locus but in different frames) are thought to be rare and, therefore, were largely neglected in t he past. In a test set of 800 viruses we found more t han 350 potent ial overlapping open reading frames of >500 bp which generate BLAST hits, indicating a possible biological fun ction . Interestingly, five overlaps with more t han 2000 bp were found, the largest may even contain triple overlaps. In order to perform the vast amount of BLAST searches required to test all detected open reading frames, we compared two clustering strategies (BLASTCLUST and k-means) and queried the database with one representative only. Our results show t hat this approach achieves a significant speed-up while retaining a high quality of the results (>99% precision compared to single queries) for both clustering methods. Future wet lab experiments are needed t o show whether the detected overlapping reading frames are biologically functional.