Tree View in T2R2
A Transform is the central concept in T2R2.
A transform is like a function that takes some input value A, some parameters and produces a value B.
Doing AI or Data Analysis in T2R2 is like in any workflow system, applying one transform after the other in a Direct Acyclic Graph.
However, in T2R2, the DAG does not have to be defined upfront. Instead, it will be deduced from T2R2 AI service.
Like in any other workflow system, users of T2R2 run transforms one after the other.
They might start with some raw data, maybe do some preprocessing, cleaning, quality checks, etc. They might continue with several AI or Data Analysis steps and eventually finish with some reporting.
Everything can be done in T2R2 but does not have to.
A worklfow is a route in a Direct Acyclic Graph of Transforms.
Here an example of a DAG-T for RNASeq.
Practically, a new transform can inherit from parent transforms and can also combine transform results, etc.
A workflow is like a tree of many transforms and those that are in parallel branches can also run in parallel on the different nodes of the cluster.
If public, a transform result can be used as a source for anybody in any new transform, hence here as well encouraging open and sound science. Data is centric.
There is no limitation on what is possible, except what a user and her company decide.
Still, each transform in the tree has all the properties of immutability and reproducibility.
In T2R2, it's also possible to clone any available transform to create a new one.
In the T2R2 UI, all relationships of a transform can be displayed starting from any transform node as one can see on the examples below.
Users can click on a node to explore its relationships further or double-click to open the transform details.
T2R2 puts the workflow engine upside-down, it works bottom-up!
Users can explore different routes without having to stick to any predefined workflows.
T2R2 AI deduces Direct Acyclic Graph paths from any data processing pipeline.
Those can then be reused, improved, adapted or completely changed. But, they are all kept forever by a smart event graph engine which ensures ultimate traceability.
Contact us to get a demo.