Skip to content

UC-4.4 - Functional Fingerprint of Samples by Pathway

Module: 4 – Functional and Genetic Profiling
Visualization type: Interactive radar (polar) plot (pathway-level KO richness for a selected sample)
Primary inputs: KEGG_Results.xlsx or KEGG_Results.csv (sample–KO–KEGG pathway associations)
Primary outputs: Pathway-level "functional fingerprint" of a selected sample

Scientific Question and Rationale

Question: What is the KO annotation fingerprint of each sample, as defined by the distribution of unique KO annotations across its metabolic pathways?

Rather than comparing samples for a single pathway, this use case focuses on characterizing one sample across all its KEGG pathways.

By summarizing, for a selected sample, the unique KO richness per pathway and representing it on a radar (polar) plot, the visualization can provide an intuitive, shape-based KO annotation fingerprint. This may reveal:

  • which pathways have particularly high or concentrated KO annotation coverage for that sample, and
  • whether the sample has annotations distributed broadly across many pathways or concentrated in a narrower set.

Data and Inputs

  • Primary data source: KEGG_Results.xlsx or KEGG_Results.csv (semicolon-delimited)
  • Key columns:
  • sample – identifier for each biological sample
  • pathname – KEGG pathway name or identifier

  • ko – KEGG Orthology (KO) identifier associated with that sample and pathway

  • User control:

  • A dropdown menu allowing selection of a single Sample (sample) for detailed profiling.

  • Output structure:

  • Axes (θ): one axis per KEGG pathway (pathname) present in the selected sample
  • Radius ®: unique KO count for each (sample, pathway) pair
  • Polygon: a closed shape connecting all pathway points, representing the sample's functional fingerprint

Analytical Workflow

  1. Sample Selection (User Input)
    The user selects a single sample from an interactive dropdown menu.

  2. All subsequent filtering and aggregation are restricted to this selected sample.

  3. Dynamic Filtering

  4. The KEGG results table KEGG_Results.xlsx or KEGG_Results.csv is loaded.

  5. The dataset is filtered to retain only rows where:

    • sample equals the selected sample, and
    • both pathname and ko are valid and non-missing.

  6. Aggregation of Pathway-Level KO Richness

  7. The filtered data is grouped by pathname.
  8. For each pathway, the number of distinct KO identifiers is computed (e.g., via nunique() on ko).

  9. This produces a set of (pathname, unique_ko_count) pairs representing the pathway-level KO richness for that sample.

  10. Rendering as Radar (Polar) Plot

  11. Each pathname is mapped to an angular coordinate (θ) around the circle.
  12. The corresponding radius ® is the unique KO count for that pathway.
  13. A closed polygon is drawn by connecting these points, optionally with markers at each vertex:
    • axes: metabolic pathways
    • radius: KO richness for the selected sample in each pathway

How to Read the Plot

  • Dropdown Menu (Sample Selection)
  • Use the menu to select the Sample whose functional fingerprint you want to inspect.

  • The radar plot recomputes and updates automatically.

  • Axes (θ – Metabolic Pathways)

  • Each radial axis represents a KEGG Pathway (pathname) for which the selected sample has at least one associated KO.

  • The set of axes forms an inventory of the pathway space encoded in that sample.

  • Radius (r – Pathway KO Richness)

  • The distance from the center along each axis is proportional to the count of unique KOs mapped to that pathway in the selected sample.

  • Higher values indicate stronger representation or greater complexity of that pathway in the sample.

  • Polygon Shape (KO Annotation Fingerprint)

  • The polygon connecting all axes encodes the overall distribution of KO annotation richness:
    • pronounced "spikes" along specific axes may indicate concentrated KO annotation coverage in those pathways
    • a more rounded, balanced shape may indicate broad and relatively even KO annotation coverage across pathways

Representative Output

The image below illustrates a representative output generated by this use case using the example dataset.

Click on the image to enlarge and explore details.

Representative output for UC-4.4

Interpretation and Key Messages

  • Pathways with Concentrated KO Annotation Coverage
  • A radar shape heavily skewed toward a subset of related pathways (e.g., several pathways within the same compound class) may indicate that the sample has concentrated KO annotation coverage in those domains.

  • Such samples may be worth prioritizing for experimental investigation of those pathways (experimental validation required to confirm functional roles).

  • KO Annotation Breadth Across Pathways

  • A more circular or evenly expanded polygon may suggest broad KO annotation coverage across many different pathways.

  • These samples may be of interest in scenarios requiring a wide range of annotated pathways to be represented.

  • Comparative Profiling Across Samples

  • By switching between samples in the dropdown, users can compare KO annotation fingerprints directly.

  • This can help identify samples with complementary or overlapping KO annotation profiles for annotation-guided hypothesis generation.

  • Link to Other BioRemPP Modules

  • When interpreted together with completeness scorecards, toxicity mapping, and regulatory alignment analyses, the KO annotation fingerprint can support annotation-based hypothesis generation and experimental planning.

Reproducibility and Assumptions

  • Input Format
    The analysis requires a semicolon-delimited KEGG results table containing at least:
  • sample,
  • pathname,

  • ko.

  • Definition of Pathway Richness

  • For each (sample, pathway) pair, pathway richness is defined as the count of unique KO identifiers mapped to that pathway.

  • Multiple occurrences of the same (sample, pathname, ko) combination do not increase the value; KOs are counted once per pathway per sample.

  • Scope and Limitations

  • The metric captures KO annotation presence rather than expression, regulation, or actual metabolic flux.
  • Radar plots are most interpretable when the number of pathways shown is moderate; in cases with very many low-richness pathways, pre-filtering (e.g., minimum KO count threshold) may be applied for clarity.

Activity diagram of the use case

Click on the image to enlarge and explore details.

Activity diagram of the use case