Distinguish drift from a selective advantage

Sampling compounds across generations

A neutral population of ten begins with five A copies. The next ten copies are drawn using probability 0.5 for A. If that generation contains six A copies, the following generation uses probability 0.6. It does not reset to the original 0.5.

The lab repeats this process for twelve generations. Every generation retains the same population size while allele counts fluctuate. Equal reproductive contribution removes the modeled selective difference, but it does not remove finite sampling.

Drift can alter allele frequencies and can be especially influential in small populations. A bottleneck changes which variants are represented among the survivors. Background: OpenStax, genetic drift.

Fixation is an endpoint under these assumptions

If all copies become A, sampling from frequency one always returns A. If A reaches zero, sampling cannot restore it. Both boundaries persist because mutation and migration are excluded. This does not mean real populations never regain variation.

The seed produces a reproducible pseudorandom trajectory. Repeating a seed checks reproducibility rather than adding independent evidence. A larger population tends to show smaller proportional fluctuations under the sampling model, but a single paired run is not guaranteed to demonstrate that tendency.

Choose comparisons before viewing outcomes

Use seeds 1, 2 and 3 at both sizes. Keep all six traces rather than selecting the most dramatic example. Compare changes from 50%, boundary events and variability across runs. Three runs illustrate variation; they are not a strong statistical study of the size effect.

Now compare these traces with the deterministic selection lesson. Both can show increasing A, but their mechanisms differ. A real population can experience both at once. To argue for selection, add reproductive evidence and a suitable neutral comparison rather than inferring advantage from one upward curve.