SUMMARY: Identifying meaningful associations between microbial communities and measured physiological or environmental variables becomes increasingly complex and computationally demanding given the continuous growth of microbiome datasets. The Coracle machine learning (ML) framework was recently developed to address this issue by integrating multiple data transformations, feature selection techniques, and ML models to yield condensed lists of features that align to target variables of interest. Further, we recently developed the UniCorP feature aggregation algorithm to identify uniquely correlated features (UNICORNs) based on the UniCor metric that iteratively enrich each taxonomic level in an automated bottom-up approach. Here we present UniCoracle, a fully automated analytical framework that integrates UniCorP's bottom-up propagation approach with a subsequent and newly developed top-down skimming (TDS) strategy, implemented with the Coracle ML framework. This combined approach leverages the inherent taxonomic structure of microbiome community data (e.g., ASVs derived from 16S rRNA gene amplicon sequencing data) to maintain predictive stability, reduce computational runtime, and identify biologically meaningful taxonomic associations. We compare the original, non-hierarchical Coracle with the TDS Coracle method and the UniCoracle approach. Evaluations across the tested datasets show that UniCoracle achieves competitive or improved predictive performance relative to both Coracle's multi-step and the TDS-based Coracle implementations and demonstrate UniCoracle's improvements in predictive accuracy over both methods. UniCoracle provides full control over feature set size and runtime, offering a streamlined and user-friendly framework for biological hypothesis generation. It identifies features (e.g., bacterial taxa) at the lowest (most specific) hierarchical level (e.g., ASV or species within a taxonomic hierarchy) that are associated with continuous target variables. AVAILABILITY: UniCoracle is freely accessible via a dedicated web server at micportal.org. The source code is open source and available on GitHub at github.com/SebastianStaab/UniCoracle.git and Zenodo at https://doi.org/10.5281/zenodo.19050205.

SUMMARY: We present Coracle, an artificial intelligence (AI) framework that can identify associations between bacterial communities and continuous variables. Coracle uses an ensemble approach of prominent feature selection methods and machine learning (ML) models to identify features, i.e. bacteria, associated with a continuous variable, e.g. host thermal tolerance. The results are aggregated into a score that incorporates the performances of the different ML models and the respective feature importance, while also considering the robustness of feature selection. Additionally, regression coefficients provide first insights into the direction of the association. We show the utility of Coracle by analyzing associations between bacterial composition data (i.e. 16S rRNA Amplicon Sequence Variants, ASVs) and coral thermal tolerance (i.e. standardized short-term heat stress-derived diagnostics). This analysis identified high-scoring bacterial taxa that were previously found associated with coral thermal tolerance. Coracle scales with feature number and performs well with hundreds to thousands of features, corresponding to the typical size of current datasets. Coracle performs best if run at a higher taxonomic level first (e.g. order or family) to identify groups of interest that can subsequently be run at the ASV level. AVAILABILITY AND IMPLEMENTATION: Coracle can be accessed via a dedicated web server that allows free and simple access: http://www.micportal.org/coracle/index. The underlying code is open-source and available via GitHub https://github.com/SebastianStaab/coracle.git.