Optimise database queries

[pverscha]

Right now, the API performs some database queries to get a list of all proteins with a specific accession number. Our API, however, requests too much information from the database, and does not split the SQL-query in smaller chunks. For one of the examples that I'm currently benchmarking, it tries to query 83k proteins at the same time causing a significant slowdown of the database.

For some endpoints, like pept2prot we also don't need to retrieve all columns from the database. By querying in a better-informed manner, we should be able to significantly speed up the process.