Create ancestry backend evidence for one backend's K sweep
Source:R/canonical_ancestry_three_backend.R
new_ancestry_backend_evidence.RdSummarizes one backend's replicate stability and fit-metric behavior across its declared K range. This does not retain raw Q matrices; those remain in the run's own evidence directory. Stability and alignment figures come from `consensus_ancestry()` run separately for each K.
Usage
new_ancestry_backend_evidence(
backend,
tool_version,
command,
k_values,
replicates,
seed,
metric_name,
stability_by_k,
provenance = list()
)Arguments
- backend
One of `admixture`, `faststructure`, or `snmf`.
- tool_version
Non-empty backend tool/package version string.
- command
Representative command or invocation template.
- k_values
Sorted, unique integer K values, each at least two.
- replicates
Positive integer replicate count per K.
- seed
Base deterministic seed.
- metric_name
Backend fit-metric name, or `NA` if unavailable.
- stability_by_k
Data frame with one row per `k_values` entry and columns `k`, `metric_mean`, `global_stability`, `mean_alignment_score`.
- provenance
Named list of additional provenance fields.