M3dical/HMOTP: HMOTP v1.0.0

Read full text Summarise with VIRAT
Authors: yihuizhou
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Fecal microbiota transplantation (FMT) is effective for recurrent Clostridioides difficile infection, but many analyses stay descriptive and do not predict individual trajectories or early response. Hierarchical Multi-Omics Trajectory Prediction (HMOTP) is a purpose-built machine learning framework for small-sample, longitudinal, multi-omics settings with high dimensionality, integration across omics, temporal structure, and interpretability. HMOTP uses hierarchical feature construction (domain knowledge), multi-level attention, and patient-specific trajectory prediction (transfer-style parameter sharing). It was evaluated on 15 recurrent CDI patients with lipidomics (397 features) and metagenomic pathways (10,634 features) at four time points over six months, using leave-one-patient-out cross-validation. HMOTP achieved strong accuracy versus single-omics and multi-omics baselines (e.g. Random Forest, Logistic Regression) and supports exploratory cross-omics and biomarker-oriented interpretation. This archive accompanies the HMOTP manuscript; the repository includes code, installation notes, and materials to reproduce reported results.

Fields of Study & Subjects

Clostridium difficile and Clostridium perfringens research Machine Learning in Healthcare Gut microbiota and health

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
26 March 2026
Language
EN