Bottom Line of Class #2:

• Currently existing algorithms of exon prediction do not work perfectly
• Databases similarity searches help a lot in detection of exons and verification of predicted exons

Other stuff:

An excellent (and readable) review on homology:

There is an important difference between similarity and homology.

Similarity is observed quantity, that might be for example expressed in percent of identity. For example, sequences might have 30% identity. On the other hand, there are no degrees of homology, as there are of similarity. The sequences are either homologous or not.

Is homology transitive?