liveBook · Manning

Model orchestration

As explained in Build AI Drug Discovery Pipelines

Model orchestration is an architectural pattern that coordinates multiple specialized models as components of a larger computational workflow.

Model orchestration coordinates several specialized models in a pipeline, passing each model’s output to the next rather than learning one shared representation. In drug discovery, this links target identification, molecular generation, virtual profiling, synthesis, and experimental testing. section 13.0 section 13.1.2

A typical orchestration workflow can connect:

  • Multi-omics and literature analysis for proposing targets. section 13.1.2
  • Molecular generation from SMILES strings or graph representations. section 13.1.2
  • Virtual ADMET prediction, followed by synthesis and experimental testing for selected candidates. section 13.1.2
  • Experimental results feeding back to refine subsequent model decisions. section 13.0

Target-to-molecule pipeline

PandaOmics can analyze multi-omics data and scientific literature to propose targets, pass validated targets to Chemistry42 for molecular generation, and apply property prediction before synthesis and testing. This reduces handoff delays and lets downstream criteria influence upstream choices. section 13.1.2

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Explains pipeline coordination and distinguishes it from shared-representation fusion, without detailing model architectures or implementation code.
Examples
Uses PandaOmics, Chemistry42, protein embeddings, molecular generation, cellular assays, and experimental feedback as workflow examples.
What you can do after
Evaluate whether a multimodal system passes outputs between models or learns a shared representation across data types.
How it connects
Connects target identification, molecular generation, ADMET prediction, virtual screening, experimental testing, and feedback loops.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build AI Drug Discovery Pipelines. The section links point to the parts of the book it draws on.