PipeAlign2

PipeAlign2 performs integrated protein family analysis by identifying sequence homologues in protein and 3D structure databases, constructing validated Multiple Alignments of Complete Sequences (MACS), defining hierarchical subfamily relationships, and clustering sequences into functional subgroups.

Key Features:

• Five-Step Analysis Pipeline: A five-step process that begins with identification of sequence homologues across protein and 3D structure databases and proceeds to define hierarchical relationships within and between subfamilies.
• Multiple Alignment of Complete Sequences (MACS): Constructs validated MACS that cluster sequences into potential functional subgroups.
• Customizable Analysis Options: Supports partial analysis runs and parameter customization for individual software modules, including refinement of existing multiple sequence alignments.
• Integration of Advanced Algorithms: Incorporates LEON for homology prediction from weak signals using residue composition and intermediate sequences, and a RASCAL-like knowledge-based refinement that focuses corrections on less reliable alignment regions within a lattice structure.
• Exploratory Data Analysis: Includes clustering techniques that automatically determine the optimal number of clusters based on density comparisons and can be applied to transcriptomic and other high-throughput datasets.

Scientific Applications:

• Homology Prediction: Aids identification of evolutionary relationships between proteins using MACS and LEON-based homology prediction methods.
• Structure and Function Analysis: Integrates sequence alignments with structural data to analyze protein family relationships at sequence and 3D-structure levels.
• High-Throughput Genome Annotation: Supports automatic genome annotation through robust alignment, clustering and subgroup delineation.
• Protein-Protein Interaction Predictions: Elucidates hierarchical relationships within subfamilies to inform potential protein-protein interaction predictions.

Methodology:

Computational steps explicitly include identification of sequence homologues in protein and 3D structure databases; construction and validation of Multiple Alignments of Complete Sequences (MACS); hierarchical subfamily definition; clustering with automatic optimal cluster number determination based on density comparisons; LEON-based homology prediction using residue composition and intermediate sequences; and RASCAL-like lattice-based alignment refinement targeting less reliable regions.