Fast LD pruning of haploid genotype matrices
BacPrune-Rust prunes a haploid genotype matrix by linkage disequilibrium (LD) threshold. Three modes are available:
--r Greedy pruning by r² (Pearson r-squared) threshold (default). --dprime Greedy pruning by |D'| threshold. --dedup Hash-based exact-duplicate removal only (O(n·v), no pairwise LD calculation).
All modes first remove exact duplicate variant columns via hashing before any threshold-based pruning. Output includes the pruned genotype matrix, a pruning summary, and a per-variant correlation-direction file.