A t-SNE Based Classification Approach to Compositional Microbiome Data

Abstract

As a data-driven dimensionality reduction and visualization tool, t-distributed stochastic neighborhood embedding (t-SNE) has been successfully applied to a variety of fields. In recent years, it has also received increasing attention for classification and regression analysis. This study presented a t-SNE based classification approach for compositional microbiome data, which enabled us to build classifiers and classify new samples in the reduced dimensional space produced by t-SNE. The Aitchison distance was employed to modify the conditional probabilities in t-SNE to account for the compositionality of microbiome data. To classify a new sample, its low-dimensional features were obtained as the weighted mean vector of its nearest neighbors in the training set. Using the low-dimensional features as input, three commonly used machine learning algorithms, logistic regression (LR), support vector machine (SVM), and decision tree (DT) were considered for classification tasks in this study. The proposed approach was applied to two disease-associated microbiome datasets, achieving better classification performance compared with the classifiers built in the original high-dimensional space. The analytic results also showed that t-SNE with Aitchison distance led to improvement of classification accuracy in both datasets. In conclusion, we have developed a t-SNE based classification approach that is suitable for compositional microbiome data and may also serve as a baseline for more complex classification models.

A More Diverse Cervical Microbiome Associates with Better Clinical Outcomes in Patients with Endometriosis: A Pilot Study

“…To compare microbial communities among groups, beta-diversity was calculated to reflect the dissimilarity between groups, including unweighted unifrac [ 44 ] and weighted unifrac distance [ 45 ]. The data in this distance matrix were further visualized by t-SNE (t-distributed stochastic neighborhood embedding), a nonlinear dimension reduction method [ 46 ], To define potential biomarkers with statistical differences among groups, metagenomic features were detected by linear discriminant analysis (LDA) effect size (LEfSe) [ 47 ]. In order to emphasize the difference of the dominant species among three samples in each taxonomic rank, the top 10 species were selected and the ternary plot was drawn based on relative abundance [ 48 ].…”

Incidence of Postoperative Pneumonia and Oral Microbiome for Patients with Cancer Operation

“…It plots the similar data points on the same map as close as possible. Therefore, tSNE is often used for microbiome analysis as a popular new ordination technique [21][22][23][24].…”

Construction of a novel signature and prediction of the immune landscape in gastric cancer based on necroptosis-related genes

“…Kaplan–Meier curves and receiver operating characteristics (ROC) curves were performed to assess the sensitivity and specificity of NRGsig. In addition, investigation was performed using principal component analysis (PCA) and T-distributed neighbor embedding (T-SNE) to analyze whether the prognostic model might properly categorize patients into two risk groups 25 . In addition, the GSE84437 cohort was used as an external validation set to confirm the model’s predictive value.…”