Taxonomy-Independent Phylogenetic Placement via De Novo Tree Binning

The patent segments the phylogenetic tree into hierarchical levels (e.g., phylum, class, order, family, genus, species) and assigns features at each level independently. This segmentation allows the system to capture fine-grained phylogenetic relationships without overwhelming complexity, as each hierarchical level can be processed and validated separately, improving precision while managing computational complexity.

The patent introduces a new dimensional approach by creating a multi-level hierarchical feature structure that extends traditional flat phylogenetic classification. By adding hierarchical depth and creating nested feature levels, the system achieves finer granularity in phylogenetic placement without linearly increasing overall system complexity, as the hierarchical structure organizes complexity in a manageable way.

The patent performs preliminary construction of a de novo phylogenetic tree from raw sequence data before any classification or analysis. This preliminary action creates a reference-free phylogenetic framework that is not biased by existing reference sets, enabling better biological meaningfulness and generalizability. The tree construction is done once and can be reused across multiple studies, managing the initial complexity investment.

The system generates its own phylogenetic features independently without relying on external reference sets. The de novo phylogenetic tree and associated features are self-generated from the input sequences themselves, making the method adaptable to any microbial community without being constrained by the scope or quality of reference databases. This self-service approach enhances generalizability across different studies and organisms.

The patent segments the feature generation process into discrete hierarchical levels, where features are extracted and validated at each level (phylum, class, order, etc.). This segmentation allows computational processing to be distributed across multiple manageable stages rather than handling all features simultaneously, reducing the burden on computational resources while maintaining fine-grained precision for reliable predictive modeling.

The patent changes the parameter of phylogenetic resolution by allowing users to adjust the granularity level of feature extraction based on computational resources available. The system can operate at different hierarchical depths (coarser or finer granularity), enabling flexible adaptation between model reliability and computational complexity requirements through parameter adjustment.