Revealing the temporal dynamics of antibiotic anomalies in the infant gut microbiome with neural jump ODEs

2025-09-30

AI Summary

Detecting abnormal microbial dynamics following antibiotic perturbations in infant gut microbiomes is challenging due to irregular, sparse longitudinal sampling. Method: We propose the first multivariate time-series anomaly detection framework based on Neural Jump Ordinary Differential Equations (NJODE), integrating conditional trajectory modeling, covariate adjustment, and path-dependent joint mean–variance inference to enable fine-grained identification and quantification of heterogeneous, multi-duration perturbations in unequally spaced longitudinal data. Results: On synthetic data, the method accurately detects diverse anomaly patterns. Applied to a real infant microbiome cohort, it reveals—for the first time—that second antibiotic exposures, prolonged courses, and post–one-year-of-age administration induce significantly more persistent dysbiosis. Moreover, it enables prospective prediction of antibiotic events, substantially outperforming baseline methods relying solely on α- or β-diversity metrics.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Time-Series/Data Streams
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Detecting anomalies in irregularly sampled multi-variate time-series is challenging, especially in data-scarce settings. Here we introduce an anomaly detection framework for irregularly sampled time-series that leverages neural jump ordinary differential equations (NJODEs). The method infers conditional mean and variance trajectories in a fully path dependent way and computes anomaly scores. On synthetic data containing jump, drift, diffusion, and noise anomalies, the framework accurately identifies diverse deviations. Applied to infant gut microbiome trajectories, it delineates the magnitude and persistence of antibiotic-induced disruptions: revealing prolonged anomalies after second antibiotic courses, extended duration treatments, and exposures during the second year of life. We further demonstrate the predictive capabilities of the inferred anomaly scores in accurately predicting antibiotic events and outperforming diversity-based baselines. Our approach accommodates unevenly spaced longitudinal observations, adjusts for static and dynamic covariates, and provides a foundation for inferring microbial anomalies induced by perturbations, offering a translational opportunity to optimize intervention regimens by minimizing microbial disruptions.

Problem

• Detecting anomalies in irregularly sampled multivariate time-series data
• Modeling antibiotic-induced disruptions in infant gut microbiome trajectories
• Providing microbial anomaly detection framework for optimizing intervention regimens

Innovation

• Neural jump ODEs detect irregular time-series anomalies
• Infers path-dependent mean and variance trajectories
• Handles uneven sampling with covariate adjustments