High performance text indexing and applications in life sciences

Eric Rivals 1, 2
2 MAB - Méthodes et Algorithmes pour la Bioinformatique
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier
Abstract : Large corpura of texts or of sequences serve as references and are interrogated through web site or programming interfaces. In life sciences, new sequencing technologies have revolutionised the acquisition of genomic sequences and triggered an exponential accumulation of reference sequences in international, public databases. Several kinds of text queries form the basic operation of programs that analyse genomic sequences. For instance, the webserver of EMBL-EBI receives 27 million queries a day. A typical sequencing experiment yields a hundred million sequencing reads-about 150 nucleotides long-each of which needs to be compared to a reference genome. To analyse such data or to mine public sequence repositories demands very efficient programs and algorithms. Only, the use of complex and specific, indexing data structures allows us to match the needs of Life sciences communities. I will present some indexing data structures that enables high performance computational analyses in genomics, and mention their pracical applications. Beyond text data, such data structures can be adapted to index other types of discrete data like trees or graphs. This will be key for the development of computational pan-genomics.
Document type :
Conference papers
Complete list of metadatas

https://hal-lirmm.ccsd.cnrs.fr/lirmm-02093450
Contributor : Eric Rivals <>
Submitted on : Monday, April 8, 2019 - 11:13:24 PM
Last modification on : Thursday, April 11, 2019 - 1:18:54 AM

File

Rivals-hpc-UK.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : lirmm-02093450, version 1

Collections

Citation

Eric Rivals. High performance text indexing and applications in life sciences. UK-France Bilateral International Meeting on High Performance Computing and Biomathematics, Feb 2019, Chicheley, United Kingdom. ⟨lirmm-02093450⟩

Share

Metrics

Record views

259

Files downloads

13