Files
obitools4/autodoc/docmd/pkg/obikmer/minimizer_utils.md
T

28 lines
1.3 KiB
Markdown
Raw Normal View History

2026-04-07 08:36:50 +02:00
# Minimizer Size Utilities in `obikmer`
This Go package provides helper functions to compute and validate the **minimizer size** `m` in k-mer-based genomic algorithms (e.g., minimizer schemes for sequence comparison or indexing).
## Core Functions
- **`DefaultMinimizerSize(k)`**
Returns a *recommended* minimizer size: `ceil(k / 2.5)`, clamped to `[1, k1]`.
→ Ensures `m` is reasonably large for uniqueness while keeping window size (`k m + 1`) manageable.
- **`MinMinimizerSize(nworkers)`**
Computes the *minimum* `m` such that there are ≥ `nworkers` distinct minimizers:
solves `4^m ≥ n_workers`, i.e., `ceil(log₄(nworkers))`.
→ Guarantees enough diversity for parallelization (e.g., hashing-based distribution across workers).
- **`ValidateMinimizerSize(m, k, nworkers)`**
Enforces constraints on `m`:
- Lower bound: ≥ `MinMinimizerSize(nworkers)` (warns & adjusts if violated)
- Hard bounds: `1 ≤ m < k`
→ Prevents invalid or inefficient parameter choices.
## Semantic Purpose
These functions ensure that minimizer-based workflows are:
- **Theoretically sound** (sufficient entropy for parallelism),
- **Practically viable** (avoiding degenerate cases like `m = 0` or `m ≥ k`),
- **User-friendly** (providing sensible defaults + clear warnings on adjustment).