TY - JOUR

T1 - Identifying interactions in omics data for clinical biomarker discovery using symbolic regression

AU - Christensen, Niels Johan

AU - Demharter, Samuel

AU - Machado, Meera

AU - Pedersen, Lykke

AU - Salvatore, Marco

AU - Stentoft-Hansen, Valdemar

AU - Iglesias, Miquel Triana

PY - 2022

Y1 - 2022

N2 - Motivation The identification of predictive biomarker signatures from omics and multi-omics data for clinical applications is an active area of research. Recent developments in assay technologies and machine learning (ML) methods have led to significant improvements in predictive performance. However, most high-performing ML methods suffer from complex architectures and lack interpretability.Results We present the application of a novel symbolic-regression-based algorithm, the QLattice, on a selection of clinical omics datasets. This approach generates parsimonious high-performing models that can both predict disease outcomes and reveal putative disease mechanisms, demonstrating the importance of selecting maximally relevant and minimally redundant features in omics-based machine-learning applications. The simplicity and high-predictive power of these biomarker signatures make them attractive tools for high-stakes applications in areas such as primary care, clinical decision-making and patient stratification.

AB - Motivation The identification of predictive biomarker signatures from omics and multi-omics data for clinical applications is an active area of research. Recent developments in assay technologies and machine learning (ML) methods have led to significant improvements in predictive performance. However, most high-performing ML methods suffer from complex architectures and lack interpretability.Results We present the application of a novel symbolic-regression-based algorithm, the QLattice, on a selection of clinical omics datasets. This approach generates parsimonious high-performing models that can both predict disease outcomes and reveal putative disease mechanisms, demonstrating the importance of selecting maximally relevant and minimally redundant features in omics-based machine-learning applications. The simplicity and high-predictive power of these biomarker signatures make them attractive tools for high-stakes applications in areas such as primary care, clinical decision-making and patient stratification.

KW - DIFFERENTIAL EXPRESSION

KW - INFERENCE

KW - MODEL

U2 - 10.1093/bioinformatics/btac405

DO - 10.1093/bioinformatics/btac405

M3 - Journal article

C2 - 35731214

VL - 38

SP - 3749

EP - 3758

JO - Bioinformatics (Online)

JF - Bioinformatics (Online)

SN - 1367-4811

IS - 15

M1 - 405

ER -