lphash

Antoine Limasset (1, 2) , Yoshihiro Shibuya (3) , Giulio Ermanno Pibiri

1 CNRS - Centre National de la Recherche Scientifique
2 CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189
3 LIGM - Laboratoire d'Informatique Gaspard-Monge

LPHash is a minimal perfect hash function (or MPHF) designed for k-mer sets where overlaps of k-1 bases between consecutive k-mers are exploited to: Reduce the space of the data structure. For example, with the right combination of parameters, LPHash is able to achive < 0.9 bits/k-mer in practice whereas the known theoretical lower-bound of a classic MPHF is 1.44 bits/k-mer and practical constructions take 2-3 bits/k-mer. Boost the evaluation speed of queries performed for consecutive k-mers. Preserve the locality of k-mers: consecutive k-mers are likely to receive consecutive hash codes. The data structure and its construction/query algorithms are described in the paper Locality-Preserving Minimal Perfect Hashing of k-mers [1].

Mots clés: Minimal perfect hash function; indexing; kmers

Domaines: Bio-informatique [q-bio.QM]

Antoine Limasset, Yoshihiro Shibuya, Giulio Ermanno Pibiri. lphash. 2024, ⟨swh:1:dir:891ec55feff3e21e72d1f8cc20c8ac06d3eedf61;origin=https://hal.archives-ouvertes.fr/hal-04448111;visit=swh:1:snp:ea894566eed31b0f93dfed739d8e5dad5706be63;anchor=swh:1:rel:8fdbe9ff5721a3774892f2e56c211adf0fad7a95;path=/⟩. ⟨hal-04448111⟩