Predictive computational phenotyping and biomarker discovery using reference-free genome comparisons

BMC Genomics

Methodology Article

Alexandre Drouin¹  Mario Marchand²  François Laviolette²  Maxime Déraspe³  Jacques Corbeil⁴  Anne-Marie Bourgault⁵  Vivian G. Loo⁵  Sébastien Giguère⁶  Michael Tyers⁶ 

[1] Department of Computer Science and Software Engineering, Université Laval, Québec, Canada;Department of Computer Science and Software Engineering, Université Laval, Québec, Canada;Big Data Research Centre, Université Laval, Québec, Canada;Department of Molecular Medicine, Université Laval, Québec, Canada;Department of Molecular Medicine, Université Laval, Québec, Canada;Big Data Research Centre, Université Laval, Québec, Canada;Division of Infectious Diseases, Departments of Medicine and Microbiology, McGill University Health Centre, Montréal, Canada;Department of Medicine, McGill University, Montréal, Canada;Institute for Research in Immunology and Cancer, Université de Montréal, Montréal, Canada;

关键词: Machine learning;    Biomarker discovery;    Antibiotic resistance;    Bacteria;    Genomics;   

DOI  :  10.1186/s12864-016-2889-6

 received in 2016-02-20, accepted in 2016-07-06,  发布年份 2016

来源: Springer

【 摘 要 】

BackgroundThe identification of genomic biomarkers is a key step towards improving diagnostic tests and therapies. We present a reference-free method for this task that relies on a k-mer representation of genomes and a machine learning algorithm that produces intelligible models. The method is computationally scalable and well-suited for whole genome sequencing studies.ResultsThe method was validated by generating models that predict the antibiotic resistance of C. difficile, M. tuberculosis, P. aeruginosa, and S. pneumoniae for 17 antibiotics. The obtained models are accurate, faithful to the biological pathways targeted by the antibiotics, and they provide insight into the process of resistance acquisition. Moreover, a theoretical analysis of the method revealed tight statistical guarantees on the accuracy of the obtained models, supporting its relevance for genomic biomarker discovery.ConclusionsOur method allows the generation of accurate and interpretable predictive models of phenotypes, which rely on a small set of genomic variations. The method is not limited to predicting antibiotic resistance in bacteria and is applicable to a variety of organisms and phenotypes. Kover, an efficient implementation of our method, is open-source and should guide biological efforts to understand a plethora of phenotypes (http://github.com/aldro61/kover/).

【 授权许可】

CC BY   
© The Author(s) 2016

【 图 表 】

Fig. 3

Fig. 2

Fig. 3

Fig. 1

Fig. 8

12951_2015_155_Article_IEq42.gif

Fig. 2

12951_2015_155_Article_IEq40.gif

Fig. 1

Fig. 3

12864_2015_2198_Article_IEq1.gif

12951_2015_155_Article_IEq39.gif

12936_2015_894_Article_IEq17.gif

Fig. 2

Fig. 1

12951_2015_155_Article_IEq37.gif

13690_2023_1196_Figb_HTML.png

13690_2023_1196_Figa_HTML.png

Fig. 1

12944_2023_1927_Article_IEq25.gif

Fig. 1

Fig. 1

12864_2016_2889_Article_IEq3.gif

Fig. 1

Fig. 2