Aquarium / Concepts

Aquarium Concepts

This document introduces the basic concepts of Aquarium. More details are given in relevant sections of the documentation.


Aquarium protocols manipulate items in the LIMS inventory. An item represents a physical entity in the lab. For instance, this diagram represents a GFP-tagged plasmid in a glycerol stock in a eppendorf tube located in a freezer.

Each item has three components representing these details: the sample, the object type, and the location. For the example in the diagram, these are

A sample is effectively the class of entity for which the item is an instance, and has a sample type. Here the sample is puc19-GFP, which has sample type of plasmid.

An item location in Aquarium is a hierarchical description of where the item can be found. For instance, the UW BIOFAB location M20.1.5.49 is a location in a box in a -20C freezer as illustrated by this diagram:


A protocol performed in Aquarium is represented as an operation. Concretely, an operation is defined by an operation type that indicates how the operation will be performed, and is defined by a protocol script that takes inputs and produces outputs. This diagram illustrates an operation type for bacterial transformation, which takes DNA and competent cells as inputs and performs a transformation to produce transformed cells.

One of the key details of this operation type is that can be given different types of inputs as long as they are consistent with the type in the operation type. In the diagram, the DNA input could be either a Maxiprep of Plasmid Library or a Miniprep of Plasmid Library. The type of the output depends on the types of the inputs.

An operation is a particular instance of an operation type with concrete inputs and outputs. This is illustrated in this diagram, which shows the bacterial transformation operation where the inputs are identified with particular items that exist in the inventory.

An operation occurs in a plan, which is a set of operations with linked inputs and outputs. The diagram shows the output of the bacterial transformation operation linked to an input of a colony PCR operation.


When plans are executed in Aquarium, similar operations are batched together as a job. These operations may come from different plans of different researchers as illustrated here where three distinct plans have shared operations.

Operations that are ready at the same time can be grouped into jobs by the manager. The operations are batched into four jobs. The manager can batch operations in to jobs as needed – in this case, the manager chose to create jobs 2 and 3 separately even though the operations have the same operation type.

Jobs are then scheduled and assigned to a technician to perform.

Operation States

After a plan is launched, the operations in the plan move through several states:

In addition, operations may have other states depending on the definition of the operation type. The most common relates to evaluation of the precondition of an operation. Each operation type has a precondition that must be true before an operation of that type can transition into pending. Most preconditions are trivially true, meaning they can always be run, but some have more complex preconditions that may fail. If the precondition of an operation fails, then the operation is put into the delayed state.

Less common is the deferred state. This arises when an operation has a predecessor operation that has an on-the-fly operation type. An operation type is marked as being on-the-fly if the number of operations of that type is used to determine the number of operations of a dependent operation type. An example is running a gel: to run a gel, you need to pour a gel. The operation type Pour Gel must be on-the-fly because it is not clear how many gels to pour until a job is formed of corresponding Run Gel operations. Because of this relationship, the Run Gel operations must be batched before the Pour Gel operations can start, and will be deferred until the Pour Gel operations complete.