Remove `convert` for biosequences

[jakobnissen]

IMO, convert should not exist for mutable types (i.e. types with object identity), because it can be called implicitly, which can lead to very weird stuff:

julia> x = dna"TAG";

julia> y = [bioseq("TAG")];

julia> push!(y, x);

julia> y[2] === x
false

[kescobo]

I saw the slack gripe on this. Doesn't feel like a huge loss, especially if we put in constructors that can do the conversion for us (eg it would be nice to be able to "convert" RNA to DNA, but it could just be with a dna sequence constructor)

[jakobnissen]

Yep, the idea is to use the constructor instead. For immutable types (say, biosymbols) we can keep the conversion.

[Bano733-code]

Instead of using the implicit convert method for mutable types like BioSequence, we could remove it to avoid confusion with object identity. A recommended approach is to use explicit constructors for conversion, e.g., DNASequence(rna_seq) or RNASequence(dna_seq). For immutable types like BioSymbols, keeping convert is fine since identity is not an issue.