Computational genome annotation

Feature type. Ideally, the reference position for elements involved in transcriptional regulation is the transcription start site (TSS). Unfortunately, the location of TSS is generally not annotated because their computer-based detection is extremely difficult and inaccurate. For most genomes, annotations only indicate the boundaries of coding sequences (CDS). For this reason, the default feature type is set to CDS. For some genomes (e.g., Drosophila, Human,...), boundaries of transcription units are also annotated. In such cases, the option mRNA can be selected as feature type, and the TSS is used as reference position for the upstream (5 ) side of genes. 

An important point to notice is the favourable conjunction of both the availability of affordable powerful computer and the arrival of fully annotated genomes from many different organisms and species, both have allowed to carry out comparative genomics and is at the source of the aforementioned methods. One can expect with newer completed genomes that these PPIs prediction methods will become more accurate and comprehensive. These approaches providing very reliable predictions already feed many PPI databases that will be discussed in the next chapter. 

---