How a "sequence" ought to comport itself.

The Bio.Behaviours.Sequence behaviour is used to define the expected functions exposed by "sequences". In general, sequences are basically a replacement for the enumerable qualities of strings in some other languages.

Because that requires defining a protocol, and that requires a struct, it makes a lot of sense to have an easy to use initializer. Thus, the new/2 method. The opts defined will depend on the type of sequence you're creating, and so the typing is left rather general here.

Because most sequences can be transcoded into other sequences (e.g. DNA -> Amino Acid), we also want to define an accessor for the module that handles that conversion. This is why the converter/0 function exists. This works together with the Bio.Behaviours.Converter, Bio.Protocols.Convertible, and Bio.Sequence.Polymer modules to create a robust conversion mechanic that can be hooked into by user defined types. For further reading on that, look at the Bio.Sequence.Polymer module docs.

The final callback, fasta_line/1, exists because this is a bioinformatics library. Sequences are pretty much always going to be written out to a fasta file, or some similar context. Defining this as a callback means that we can make it easier for your types to be given directly to the Bio.IO.Fasta module for writing. Eventually, I'd probably like to come up with a more general dump style mechanic. But this'll do for pre-alpha.