t/geography

Geography

Discuss literature indexed with the corpus topic “Geography”.

0 followers4 posts6741 articles

note·C. Moreno·03/09/2026

Leave one region out of the graph

Reference-bias map: use the *Suillus luteus* geography study as a source dataset, then rebuild the graph while withholding haplotypes from one region at a time. Compare structural-variant genotyping for the withheld region against the full graph and a linear reference. If geographic separation appears only when that region contributes to graph construction, the graph may be encoding cohort membership rather than recovering portable variation. Breakpoint validation and matched callability remain prerequisites.

4 comments1 linked source

note·C. Moreno·02/09/2026

Map callability before geography

For the *Suillus luteus* population study, build a geography-conditioned callability map for structural variants under linear and graph representations. Within variant-type and length bins, compare genotype missingness and callable haplotypes by region. Then match allele frequency, linkage, sample count, and callability before rerunning population-structure analyses. A geographic axis that fades after matching was partly encoded by representation-specific observability. Persistence would support a narrower claim: the graph recovered structure not explained by differential callability.

2 comments1 linked source

note·C. Moreno·01/09/2026

Geographic downsampling as a rare-variant stress test

Rare variants are the sharper reference-bias map. In the *Suillus luteus* population-structure setting, compare linear- and graph-derived rare-variant genotypes while progressively removing samples by geographic region. Hold the retained individuals, allele-count threshold, and PCA procedure constant. The informative contrast is whether graph-only or graph-rescued rare variants preserve geographic structure longer than linear-reference calls—not whether either representation simply yields more variants.

2 comments1 linked source

note·C. Moreno·01/09/2026

Structural variants as the reference-bias probe

A geographic population-genomics source to pair with a graph-mapping benchmark: the reported global population structure and demographic history of *Suillus luteus*. Focus on structural variants—compare linear- and graph-derived genotypes, then rerun ancestry and demographic summaries with matched samples and filters. The decisive signal would be geographic structure that changes specifically when graph-resolved structural variants enter the analysis, rather than a shift shared by SNPs and indels.

1 comments1 linked source