How small can a governed synthetic pool be? A minimum viable pool study with a pre-registered fidelity statistic
Description
This study determines the smallest governed synthetic genetic pool that continues to track its source population under criteria fixed in advance. Governed (lambda = 0.55) and neutral branches were propagated on shared founders at pool sizes of 50 to 800 diploid founders, over six generations and 30 replicates, on the five-locus MC1R reference panel seeded from gnomAD v4 non-Finnish European frequencies. In the frozen tier, fidelity was scored by an ensemble Fisher z statistic on the correlation between governance-induced allele frequency shifts and the panel weights. The frozen run is void: the sanity criterion failed at N = 400, fidelity failed at every size, and heterozygosity retention passed at every size. A dated amendment (A-S1) corrected the fidelity predictor to the breeder's equation form and added a locus-loss component, on fresh replicates. Under the amended tier, fidelity and retention passed at every size and locus loss was zero at N = 200 and above. The minimum viable pool under Amendment A-S1 is N = 200. Both tiers are reported in full, and the frozen verdict stands unaltered.
Files
fig_s1_retention.png
Files
(417.2 kB)
| Name | Size | |
|---|---|---|
|
md5:9562799cd8285f397f4cab06a2de8413
|
83.8 kB | Preview Download |
|
md5:d33bb73353850784ee7e534f6ab46be9
|
68.7 kB | Preview Download |
|
md5:a611dc073bedbb5a0001fbf68a31a09b
|
65.9 kB | Preview Download |
|
md5:39e5efd5595152b2c26883dbe534a6e7
|
86.1 kB | Preview Download |
|
md5:325403fb8feb6d26c447562f7d2c9ce5
|
112.6 kB | Preview Download |