User Workflow
The typical Pheno-Ranker workflow has four stages: prepare input data, run a comparison, inspect or visualize the outputs, and optionally use companion utilities.
Start with your data format, choose the comparison mode, then follow the output to the analysis or utility page that matches your task.
Prepare data
Start from BFF, PXF, configured generic JSON/YAML, or tabular data converted with csv2pheno-ranker.
Compare records
Run cohort mode for all-vs-all comparisons or patient mode to rank matches against a target patient.
Analyze outputs
Inspect rankings, matrices, sparse Matrix Market files, graph exports, or downstream visualizations.
Use utilities
Run companion CLI tools to simulate data, import CSV files, plot BFF/PXF summaries, or encode vectors as QR codes.
1. Prepare Data
Use BFF or PXF directly when your data already follows GA4GH-oriented formats. For custom JSON or YAML, provide a configuration file that defines the record identifier and comparable terms. For tabular data, use csv2pheno-ranker to create both the JSON input and the configuration file.
2. Compare Records
Run cohort mode for all-vs-all comparisons within or across cohorts, or patient mode to rank a reference cohort against a target profile. Both workflows use flattened records and one-hot encoded binary vectors.
3. Analyze Outputs
The main outputs are plain text files that can be inspected directly or used downstream. Cohort mode produces a pairwise matrix suitable for clustering, multidimensional scaling, or graph analytics. Patient mode produces ranked matches with Hamming distance, Jaccard similarity, Z-scores, p-values, and overlap statistics.
4. Use Companion Utilities
The toolkit also includes companion command-line utilities to simulate BFF/PXF data, convert CSV files, plot BFF/PXF summaries, and encode Pheno-Ranker vectors as QR codes. These utilities are useful for testing installations, preparing non-GA4GH data, and building reproducible examples.