MIMP mutations mapping - mutli PTM sites issue

[Michał Krassowski]

When reviewing the MIMP mutations importing code I've noticed that one-to-one correspondence is required to create a MIMP mutation associated with a site. This hoverer means that sites with multiple known PTM modifications are missed. This may have affected up to 200 or 0.05% of (acetylation + phosphorylation) sites in v2017, and becomes more pronounced in v2019 after addition of glycosylation sites which include further 300 sites in (glycosylation + phosphorylation) overlap.

Instead, we should be matching the site by type of the motif that the mutation breaks (I mean, there will be more types - i.e. types other than phosphorylation - if we include the glycosylation sites motif development from https://github.com/reimandlab/rmimp/tree/refactored).