Create model

There are several nodes: Variety. Each node is connected with the nodes :Control (no treatment applied), :P1R - one treatment is applied and :P2B another treatment is applied. Each node ( :Control, :P1R, :P2B) is connected to the :Protein node by the relation :CONTAINS (contains a certain percentage of protein).
The protein content of each variety must be determined.
Is such a variant of the subgraph possible?
How can it be determined to which variety the value in the Protein node belongs?

A couple questions to help understand your donain:

  1. What is the formula to determine the protein content for one variety?
  2. Why do you have three CONTAINS relationships connecting each treatment and control node to the protein node?

For each variety there is a control group and a group with the applied treatment.
Control group of Variety 1 (Clavera) contains x% protein (Mean value);
The second group of the same variety treated with the first Compound (P1R) contains x% protein;
The third group of the same variety is treated with P2B and contains x% protein.
And the same goes for all varieties.
Based on this model, the influence of the treatment on the protein content will be assessed. The Pearson Similarity algorithm could be used.

Per your model, you need to store some info about Variety node as a property in CONTAINS relationship:
create (pl)-[rel:CONTAINS {percentages: 33.56, paid: 4}]->(prot)
With this you find the protein percentages for each Variety,