Correlations & Analysis

Biomarkers Γ— WHOOP sleep/recovery Β· Strava training load Β· life covariates β€” 2013–2026

Key Insights

Top Correlations β€” Age-Adjusted (partial r)
Ranked by partial r after removing the shared linear trend with age from both the covariate and the biomarker. This answers "does this metric predict the biomarker beyond what aging alone explains?" β€” the more meaningful question given that many things change with age simultaneously. n β‰₯ 5 required.
Raw Correlations (unadjusted, for reference)
Pearson r without age-adjustment. Ranked by |r|. Anything that also has a large age trend may be partially or fully confounded.

Scatter Plots β€” Top Pairs (raw)

Each point is one blood draw (faint = early, solid = recent). Regression line shown. Age not removed from axes.

Biomarker Γ— All Covariates
Every biomarker row vs. all covariates. Raw r: plain Pearson correlation. Age-adjusted partial r: residuals after regressing both variables on age β€” measures the relationship independent of aging. Age-adjusted is blank for the age column itself and where n < 5. Scroll right for all columns.
βˆ’1+1Β· blank = not enough shared draws
Top Biomarker–Biomarker Correlations
Pearson r between all biomarker pairs on shared blood-draw dates (β‰₯4 required). Top 20 by |r|.

Scatter Plots β€” Key Biomarker Pairs

Each point = one blood draw. Faint (early) β†’ solid (recent).

Wearable Metrics Over Blood-Draw Dates

Rolling averages of WHOOP and Strava signals plotted at each blood-draw date. Life-event markers shown as dashed vertical lines where they fall in range.

Estimated VOβ‚‚max β€” Full History (2011–2026)

VOβ‚‚max estimated series
Lab measurements (2 points) anchored to estimated series from: VDOT (Jack Daniels formula from race times) and FTP w/kg (Coggan formula, calibrated to lab values). Non-lab estimates are scaled so they match the lab anchors within Β±180 days. Each method shown in a different colour. Blood-draw dates marked.

Coverage Summary