Skip to content

From assertions to synthesis

EukTrait preserves biological knowledge as a collection of assertions, each tied to a taxon or material, a feature, a trait, and a source.
But scientists rarely want to read raw assertions one by one. They want synthesized information: summaries of what is known about a species, trait distributions, or ecological patterns.

This page describes how assertions can be aggregated and interpreted conceptually, while still respecting provenance, context, and variability.

The challenge of synthesis

Biological knowledge is rarely simple:

  • Traits vary between life stages, environments, and strains.
  • Different sources may make conflicting claims.
  • Some traits are observed directly, others are inferred.

Simply collapsing assertions into a table risks losing context, erasing uncertainty, and conflating observations with generalizations.

EukTrait addresses this by keeping assertions as the canonical record, while providing principles for deriving synthesized views.

Conceptual steps to synthesis

While implementation may vary, synthesis generally involves three stages:

1. Selecting the scope

Decide which assertions to consider:

  • Taxon-level or material-level?
  • Specific life stages, environments, or experimental methods?
  • Direct observations only, or including inferences?

Using qualifiers to filter assertions ensures that the synthesis reflects exactly the intended context.

2. Grouping and aggregation

Assertions are grouped by:

  • taxon or material,
  • feature and trait,
  • optionally, other contextual qualifiers (life stage, evidence type).

Aggregation may include:

  • counting how many sources report a trait,
  • identifying the range of values observed,
  • summarizing conflicting assertions with context intact.

Example (conceptual):

Trait: feeding_mechanism
Taxon: Rictus lutensis

Observed:
  - phagotrophy (material RCC299, light microscopy)
  - phagotrophy (material RCC300, electron microscopy)
Inferred:
  - chemotrophy (species-wide, inferred from morphology)

This keeps both the observations and inferences visible.

3. Interpreting patterns

Once aggregated, synthesized views can answer questions like:

  • What feeding strategies are reported for this species, and under what conditions?
  • Which traits are variable between materials or life stages?
  • Where do sources agree or conflict?

Importantly, no assertion is discarded. Users can always drill down to the original claim.

Best practices for synthesis

  1. Preserve provenance.
    Always retain links to the sources supporting each assertion.

  2. Respect context.
    Life stage, environment, and evidence method are not optional; they shape interpretation.

  3. Distinguish observation from inference.
    Aggregated views should mark which claims were measured directly versus inferred.

  4. Do not overgeneralize.
    Avoid assuming that a trait observed in one material applies to the entire taxon unless explicitly supported.

  5. Document decisions.
    Any filtering, aggregation rules, or weighting applied during synthesis should be transparent and reproducible.

Conceptual example: trophic traits

Taxon: Rictus lutensis

Feature: organism
Trait: feeding_mechanism

Aggregated summary:
  - phagotrophy: observed in RCC299, RCC300 (active stage, microscopy)
  - chemotrophy: inferred from morphology (species-wide)

Here, both observed and inferred assertions coexist. Users can see which traits are directly supported, and which are inferred.

Why this matters

Synthesis is how EukTrait becomes useful for comparative biology:

  • It allows identifying patterns across taxa or materials.
  • It preserves complexity rather than flattening it.
  • It keeps evidence explicit, allowing reinterpretation as new data appear.

Even without fully automated tools, thinking in terms of assertions and synthesis ensures that all trait knowledge is traceable, transparent, and contextually meaningful.

Looking forward

Future implementations may include:

  • automated aggregation pipelines,
  • trait dashboards for taxa or clades,
  • visualizations of variability and conflict,
  • and programmatic interfaces for comparative analysis.

Until then, the principles described here serve as a guide for manual synthesis and careful curation, ensuring that EukTrait remains both flexible and robust.