Polygenic Embryo Selection: Selection Differential Simulator

Embryo Batch

Batch size N 8

Selected per batch k 1

Trait heritability h² 0.50

Selection goal

Live Readouts

Selection differential S

0.00 σ

Predicted response R = h²S

0.00 σ

Selected group population percentile

50th

Batch (polygenic z-score)

How it works

Each embryo gets a polygenic index (PGI) for one trait — a standardized z-score (mean 0, SD 1 in the reference population) built by summing thousands of small trait-associated variants, exactly as real preimplantation polygenic screening reports it. Siblings from the same two parents scatter around the mid-parent value purely from independent assortment at meiosis, which this batch approximates as z ~ N(0,1).

Selection differential:  S = mean(selected z) − mean(batch z)
Breeder's equation:      R = h² · S
Selected percentile:     Φ(mean selected z) · 100

Selecting the top-k embryos out of the batch creates a selection differential S — how far above (or, for a risk index, below) the batch average the chosen group sits. The breeder's equation R = h²S, the same formula that predicts a trait's response to artificial selection in agriculture, converts that differential into the expected phenotypic shift: only the heritable fraction h² of the statistical difference actually shows up as a real trait change, the rest is noise the index can't capture.

N / k sliders — how many embryos are in the batch and how many are chosen as "top" by the polygenic index.
h² slider — the trait's heritability; most complex traits (height ≈0.8, common disease risk often 0.2–0.5) sit well below 1, so R is always smaller than S.
Selection goal — maximize a trait index or minimize a disease-risk index; the math (and the ethical debate) is identical either way.
This is the exact quantitative-genetics argument behind the "designer baby" and polygenic embryo screening debate: with realistic sibling counts and real-trait heritabilities, R stays only a fraction of a standard deviation — a small, probabilistic nudge, not a chosen trait.

Population distribution N(0,1) Selected embryo Not selected

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.