Phenofhy

Profiling a Phenotype ​

This tutorial explains what profile.py produces and how to use it.

What phenotype_profile() generates ​

phenotype_profile() creates a multi panel figure and summary metadata for a single phenotype. The figure is designed as a one page report with the following sections:

  • Metadata: field name, entity, type, and units (if available in a data dictionary).
  • Demographic summary table (counts and percent by sex, plus age summary stats).
  • Phenotype completeness (missing vs non-missing for the target phenotype).
  • Distribution by sex (histogram for continuous, bar chart for categorical).
  • Mean or proportion trends by age and body measurements (binned by age, BMI, or height).

The function also returns a dictionary with the phenotype type, sample flow, and demographics table, and can optionally save the figure to a PDF.

Example ​

python
from phenofhy.profile import phenotype_profile

fig, meta = phenotype_profile(
    df,
    phenotype="clinic_measurements.heart_first_rate",
    age_bin_width=2,
    body_bin_width=1,
    min_n_per_bin=30,
    error_mode="se",
    output="outputs/profile_heart_rate.pdf",
)

Notes ​

  • For continuous phenotypes, trend plots show mean with SE by default (use error_mode="sd" for SD).
  • For binary phenotypes, trend plots show proportion with SE by default.
  • Categorical phenotypes do not show trend plots.
  • The data dictionary lookup uses the first matching *data_dictionary.csv in metadata_dir.