Interpreting k-mer–based signatures for antibiotic resistance prediction. Issue 10 (17th October 2020)

• Record Type: Journal Article
• Authors: Jaillard, Magali; Palmieri, Mattia; van Belkum, Alex; Mahé, Pierre
• Is Part Of: GigaScience. Volume 9:Issue 10(2020)
• Publication Date: 2020-10-17
• Subjects: antibiotic resistance -- supervised machine learning -- k-mer -- de Bruijn graph; Information storage and retrieval systems -- Research -- Periodicals; Biology -- Research -- Periodicals; Medical sciences -- Research -- Periodicals; Database management -- Periodicals; 570.285
• DOI: 10.1093/gigascience/giaa110
• Languages: English
• ISSNs: 2047-217X

Abstract

Abstract: Background: Recent years have witnessed the development of several k -mer–based approaches aiming to predict phenotypic traits of bacteria on the basis of their whole-genome sequences. While often convincing in terms of predictive performance, the underlying models are in general not straightforward to interpret, the interplay between the actual genetic determinant and its translation as k -mers being generally hard to decipher. Results: We propose a simple and computationally efficient strategy allowing one to cope with the high correlation inherent to k -mer–based representations in supervised machine learning models, leading to concise and easily interpretable signatures. We demonstrate the benefit of this approach on the task of predicting the antibiotic resistance profile of a Klebsiella pneumoniae strain from its genome, where our method leads to signatures defined as weighted linear combinations of genetic elements that can easily be identified as genuine antibiotic resistance determinants, with state-of-the-art predictive performance. Conclusions: By enhancing the interpretability of genomic k -mer–based antibiotic resistance prediction models, our approach improves their clinical utility and hence will facilitate their adoption in routine diagnostics by clinicians and microbiologists. While antibiotic resistance was the motivating application, the method is generic and can be transposed to any other bacterial trait. An R package implementing our method is available at https://gitlab.com/biomerieux-data-science/clustlasso .