Caretta

Caretta performs multiple protein structure alignment and extracts aligned structural feature matrices to enable comparative analysis and machine-learning–based prediction of functional properties in homologous but sequence-divergent protein families.

Key Features:

• Multiple Structure Alignment: Generates robust multiple structure alignments that capture conserved structural features across homologous proteins with significant sequence divergence.
• Higher Coverage and Accuracy: Produces alignments reported to have increased coverage and improved accuracy relative to existing state-of-the-art methods.
• Feature Matrix Output: Outputs an aligned structure feature matrix for input protein structures suitable for downstream computational analyses.
• Machine-learning Feature Extraction: Extracts structural features intended for integration into supervised and unsupervised machine-learning workflows.

Scientific Applications:

• Homology modeling integration: Enables incorporation of both sequence and structural features into predictive models used in homology modeling workflows.
• Conformational-state prediction for CDKs: Supports derivation of structural insights to model and predict conformational states of cyclin-dependent kinases (CDKs).

Methodology:

Performs multiple protein structure alignment and feature extraction, producing aligned structure feature matrices for the input protein structures for use in downstream analyses.