Improved Prediction of Bacterial Genotype-Phenotype Associations Using Interpretable Pangenome-Spanning Regressions

Abstract

<jats:title>ABSTRACT</jats:title> <jats:p>Discovery of genetic variants underlying bacterial phenotypes and the prediction of phenotypes such as antibiotic resistance are fundamental tasks in bacterial genomics. Genome-wide association study (GWAS) methods have been applied to study these relations, but the plastic nature of bacterial genomes and the clonal structure of bacterial populations creates challenges. We introduce an alignment-free method which finds sets of loci associated with bacterial phenotypes, quantifies the total effect of genetics on the phenotype, and allows accurate phenotype prediction, all within a single computationally scalable joint modeling framework. Genetic variants covering the entire pangenome are compactly represented by extended DNA sequence words known as unitigs, and model fitting is achieved using elastic net penalization, an extension of standard multiple regression. Using an extensive set of state-of-the-art bacterial population genomic data sets, we demonstrate that our approach performs accurate phenotype prediction, comparable to popular machine learning methods, while retaining both interpretability and computational efficiency. Compared to those of previous approaches, which test each genotype-phenotype association separately for each variant and apply a significance threshold, the variants selected by our joint modeling approach overlap substantially.</jats:p> <jats:p><jats:bold>IMPORTANCE</jats:bold> Being able to identify the genetic variants responsible for specific bacterial phenotypes has been the goal of bacterial genetics since its inception and is fundamental to our current level of understanding of bacteria. This identification has been based primarily on painstaking experimentation, but the availability of large data sets of whole genomes with associated phenotype metadata promises to revolutionize this approach, not least for important clinical phenotypes that are not amenable to laboratory analysis. These models of phenotype-genotype association can in the future be used for rapid prediction of clinically important phenotypes such as antibiotic resistance and virulence by rapid-turnaround or point-of-care tests. However, despite much effort being put into adapting genome-wide association study (GWAS) approaches to cope with bacterium-specific problems, such as strong population structure and horizontal gene exchange, current approaches are not yet optimal. We describe a method that advances methodology for both association and generation of portable prediction models.</jats:p>

Interpreting k-mer–based signatures for antibiotic resistance prediction

“…The elastic-net penalty combines the lasso and the ridge penalties, which leads to sparse models with a grouping mechanism: correlated features tend to be selected together [ 46 ]. This approach was recently shown to be efficient in the context of bacterial genome-wide association studies (GWAS), providing increased statistical power for the identification of genotype-phenotype associations and accurate prediction rules [ 47 ]. As we demonstrate in Supplementary Section S9 , however, it remains limited in its ability to provide interpretable predictive signatures, for several reasons.…”

Boosting heritability: estimating the genetic component of phenotypic variation with multiple sample splitting

“…The tuning parameter > 0 controls the overall strength of the penalty and we use 10-fold crossvalidation to choose suitable value for . Elastic net approach is implemented in the software 'pyseer' [36,37] focusing on GWAS for bacterial data.…”

Learning From Limited Data: Towards Best Practice Techniques for Antimicrobial Resistance Prediction From Whole Genome Sequencing Data

“…We selected three machine learning algorithms for prediction of antimicrobial resistance from WGS data represented as nucleotide k-mer profiles: extreme gradient boosting (XGB) ( Chen and Guestrin, 2016 ), elastic net regularized logistic regression (ENLR) ( Friedman et al., 2010 ), and set covering machine (SCM) ( Marchand and Shawe-taylor, 2000 ). All selected algorithms were recently reported to perform well on the WGS-AST task ( Aun et al., 2018 ; Nguyen et al., 2018a ; Drouin et al., 2019 ; Ferreira et al., 2020 ; Lees et al., 2020 ).…”