ISA vs MIAPPE Comparison

ISA and MIAPPE were developed by different communities for different purposes, but they share common concepts. Both have Investigation and Study as top-level containers, both track experimental factors, and both produce data files. However, they model the experimental process differently.

ISA focuses on the transformations that occur during an experiment - how biological material is collected, processed, and measured. It uses a directed graph of Processes to capture provenance.

MIAPPE focuses on what is being observed and measured. It uses ObservationUnits as the central concept, with Events and Environment providing context for when and where measurements were taken.

When working with plant experiments that involve both phenotyping and molecular assays (e.g., measuring plant height and also doing RNA-seq), consider using the profile merge feature to create a custom combined profile.

Entity Comparison

flowchart TB
    subgraph shared["Shared Concepts"]
        direction TB
        INV[Investigation]
        STU[Study]
        PER[Person/Contact]
        SAM[Sample]
        FAC[Factor/StudyFactor]
        FV[FactorValue]
        DF[DataFile]
    end

    subgraph isa["ISA-Specific"]
        direction TB
        ASS[Assay]
        PROT[Protocol]
        PP[ProtocolParameter]
        SRC[Source]
        EXT[Extract]
        LEXT[LabeledExtract]
        PROC[Process]
        OA[OntologyAnnotation]
        OS[OntologySource]
        CHAR[Characteristic]
        PV[ParameterValue]
        PUB_ISA[Publication]
        COM[Comment]
    end

    subgraph miappe["MIAPPE-Specific"]
        direction TB
        BM[BiologicalMaterial]
        OU[ObservationUnit]
        OV[ObservedVariable]
        EVT[Event]
        ENV[Environment]
        LOC[Location]
        MS[MaterialSource]
    end

    INV --> STU
    INV --> PER
    STU --> FAC
    FAC --> FV

    STU --> ASS
    ASS --> DF
    ASS --> PROC
    STU --> PROT
    PROT --> PP
    STU --> SRC
    SRC -->|Process| SAM
    SAM -->|Process| EXT
    EXT -->|Process| LEXT
    PROC --> PV
    INV --> PUB_ISA

    STU --> BM
    STU --> OU
    STU --> OV
    STU --> EVT
    STU --> ENV
    OU --> SAM
    BM --> MS
    STU -.-> LOC

    classDef shared fill:#e8f5e9,stroke:#4caf50,stroke-width:2px
    classDef isa fill:#e3f2fd,stroke:#2196f3,stroke-width:2px
    classDef miappe fill:#fff3e0,stroke:#ff9800,stroke-width:2px

    class INV,STU,PER,SAM,FAC,FV,DF shared
    class ASS,PROT,PP,SRC,EXT,LEXT,PROC,OA,OS,CHAR,PV,PUB_ISA,COM isa
    class BM,OU,OV,EVT,ENV,LOC,MS miappe

How They Model Experiments

ISA answers: "How was this data produced?"

An ISA experiment is a chain of transformations. You start with a Source (a patient, a plant), collect a Sample, extract molecules (Extract), label them for measurement (LabeledExtract), and run an assay that produces data. Each transformation is a Process that references a Protocol describing the method used.

MIAPPE answers: "What was measured and under what conditions?"

A MIAPPE experiment defines ObservationUnits (the things being measured - plants, plots) and ObservedVariables (the traits being measured - height, yield). Events record what happened (planting, watering) and Environment records conditions (temperature, humidity).

Limitations

ISA lacks entities for environmental context (temperature, humidity) and temporal events (planting, harvest). Phenotypic trait definitions require workarounds through Characteristics or Comments. No Location entity exists for geographic context, and no ObservedVariable structure (trait/method/scale) for standardized measurement definitions. The format was designed for omics workflows, not field trials.

MIAPPE lacks process-level provenance. Molecular workflows (extraction, labeling, sequencing) cannot be captured. Protocol documentation is informal compared to ISA's structured ProtocolParameter system. The format is plant-specific by design (BiologicalMaterial assumes organism/genus/species). No OntologySource management reduces semantic rigor. No Assay concept exists to group measurements by technology.

Both formats cannot represent experiments that combine phenotyping and molecular assays within a single coherent model.

References

Standard	Specification	GitHub
ISA	https://isa-specs.readthedocs.io/	https://github.com/ISA-tools/isa-api
MIAPPE	https://www.miappe.org/	https://github.com/MIAPPE/MIAPPE