Hierarchical structural component model for pathway analysis of multinomial phenotypes.

backgroundMany statistical methods for pathway analysis have been used to identify novel pathways from biomarkers associated with a certain disease. However, most of these methods are based on single pathway analysis and do not consider multiple pathways simultaneously. To address this issue, a hierarchical structural component model (HisCoM) was developed, which takes into account all pathways at the same time, as well as takes into consideration the correlations among them. HisCoM has been successfully applied to the analysis of continuous, count, and binary phenotypes.

objectiveIn this study, our goal is to propose HisCoM-Categ by extending HisCoM for pathway analysis for both nominal or ordinal multinomial phenotypes, when the phenotypes have more than two possible unordered or ordered discrete categories.

methodsThe foundation of the proposed HisCoM-Categ is the multivariate extension of generalized linear models. Specifically, HisCoM-Categ accounts for the hierarchical structure of biomarkers and pathways, as well as the correlations that exist among pathways.

resultsThrough the simulation study, HisCoM-Categ was shown to have higher power compared to the other existing methods. In addition, HisCoM-Categ was illustrated with two different omics datasets, including metabolomic, and metagenomic datasets. HisCoM-Categ for ordinal multinomial phenotypes was illustrated by the metabolomic and metagenomic datasets. Those applications demonstrated that HisCoM-Categ successfully identified the well-known pathways that are associated with multinomial phenotypes.

conclusionsThe current study proposes a novel pathway analysis method HisCoM-Categ to identify pathways that have been associated with multinomial phenotypes.