HMMERCTTER

HMMERCTTER classifies protein superfamily sequences by combining high-quality superfamily phylogeny-based clustering with HMMER profile construction and cluster-specific score thresholds to enable precise subfamily assignment.

Key Features:

• Integration with HMMER: Builds HMMER profiles from clustered training sequences and uses HMMER score thresholds for classification.
• High-quality phylogeny utilization: Leverages high-quality superfamily phylogenies to cluster training sequences prior to profile construction.
• Cluster-specific thresholds: Generates a specific inclusion cut-off threshold for each cluster to distinguish subfamily members.
• Iterative sequence inclusion: Iteratively incorporates novel sequences into groups and updates the corresponding HMMER profiles to refine detection.
• Precision and recall performance: Demonstrated case-study performance includes 100% precision with over 95% recall on large protein superfamily datasets.

Scientific Applications:

• Sequence function assignation: Supports decision-making for functional assignment in regions of ambiguous sequence similarity.
• Subfamily identification: Facilitates precise identification of subfamilies within protein superfamilies.
• Functional genomics and proteomics: Aids analyses that require accurate subfamily-level classification for downstream genomics and proteomics studies.

Methodology:

Use a high-quality superfamily phylogeny to cluster training sequences, build HMMER profiles for each cluster and determine cluster-specific score thresholds that achieve 100% precision and recall for detecting cluster members, apply those profiles and thresholds to screen target datasets, and iteratively include novel sequences while updating profiles.