How to integrate with PickleType

Nobody has claimed this yet.

I want to pass a list to the schema which uses SQLAlchemy's `PickleType` column.
The json file will looks like this:
```
json_data = {uuid: '123456...', coords: [[x1, y1], [x2, y2]]}
```
But looks like `ma.SQLAlchemyAutoSchema` treat `PickleType` column as string so if I do `schema.load(json_data)` it will say `coords` is `not a valid string`.
I use some trick like `str(json_data(coords)` before `schema.load(json_data)` and `eval(json_data(coords)` after. But when I do `schema.dump(model)`, it will still convert `coords` to string. And I have to `eval` it again.
Is there better way to to integrate with PickleType? Or is there any way to skip the validation for specific column.

Start by reproducing the reported load and dump behavior through ma.SQLAlchemyAutoSchema with a SQLAlchemy PickleType column and the coords list shown in the issue. The payload names no files or tests; done means establishing whether PickleType values can round-trip without string conversion or eval, or documenting the supported way to skip or customize validation.

Written by the indexing model from the issue text.