Litcius/Paper detail

Interpretable Log Contrasts for the Classification of Health Biomarkers: a New Approach to Balance Selection

Thomas P. Quinn, Ionas Erb

2020mSystems32 citationsDOIOpen Access PDF

Abstract

High-throughput sequencing provides an easy and cost-effective way to measure the relative abundance of bacteria in any environmental or biological sample. When these samples come from humans, the microbiome signatures can act as biomarkers for disease prediction. However, because bacterial abundance is measured as a composition, the data have unique properties that make conventional analyses inappropriate. To overcome this, analysts often use cumbersome normalizations. This article proposes an alternative method that identifies pairs and trios of bacteria whose stoichiometric presence can differentiate between diseased and nondiseased samples. By using interpretable log contrasts called balances, we developed an entirely normalization-free classification procedure that reduces the feature space and improves the interpretability, without sacrificing classifier performance.

Topics & Concepts

InterpretabilityNormalization (sociology)Classifier (UML)Computer scienceFeature selectionArtificial intelligenceRelative species abundanceMicrobiomeMachine learningData miningPattern recognition (psychology)Computational biologyBiologyAbundance (ecology)BioinformaticsEcologyAnthropologySociologyMetabolomics and Mass Spectrometry StudiesGene expression and cancer classificationBiomedical Text Mining and Ontologies
Interpretable Log Contrasts for the Classification of Health Biomarkers: a New Approach to Balance Selection | Litcius