English

Unclustered BWTs of any Length over Non-Binary Alphabets

Discrete Mathematics 2025-08-29 v1 Data Structures and Algorithms Formal Languages and Automata Theory Combinatorics

Abstract

We prove that for every integer n>0n > 0 and for every alphabet Σk\Sigma_k of size k3k \geq 3, there exists a necklace of length nn whose Burrows-Wheeler Transform (BWT) is completely unclustered, i.e., it consists of exactly nn runs with no two consecutive equal symbols. These words represent the worst-case behavior of the BWT for clustering, since the number of BWT runs is maximized. We also establish a lower bound on their number. This contrasts with the binary case, where the existence of infinitely many completely unclustered BWTs is still an open problem, related to Artin's conjecture on primitive roots.

Keywords

Cite

@article{arxiv.2508.20879,
  title  = {Unclustered BWTs of any Length over Non-Binary Alphabets},
  author = {Gabriele Fici and Estéban Gabory and Giuseppe Romana and Marinella Sciortino},
  journal= {arXiv preprint arXiv:2508.20879},
  year   = {2025}
}