Ideal way to store a list of transforms and their collective parents? Details inside.

I have my own collision detection system, and part of this system is a 'Mover' class attached to an empty transform with many Collider2Ds as child objects.

I often need to check a collision to see if it has hit a Mover type, and access the Mover class; using GetComponentInParent each time would be inefficient so I'd like to cache the structure somehow.

So here are some things I've considered, but I don't really have enough knowledge to make a decision, which is why I'm here! It's possible I'm missing some obvious answer.

Idea 1:- Keep a single List that every Mover registers itself into when spawning. Then within each mover have a HashSet. Populate the HashSet when the Mover spawns using GetComponentsInChildren. When testing a raycast collision, iterate through the Mover list and check the Hashset for the hit.collider

Idea 2:- Keep a single Dictionary. Populate the HashSet when the Mover spawns using GetComponentsInChildren. When testing a raycast collision, search the dictionary using hit.collider as the key.

Alternatively I could place all my Mover objects into their own physics layer and query the layer - though I still need an efficient way to access the Mover class after confirmation.

(edit) Clarification of the transform structure,

 Root
     Mover
         Collider2D
         Collider2D
     Mover    <------.  want this
         Collider2D   \
         Collider2D <------ Hit this        
         Collider2D    
     Mover
         Collider2D
     etc