TRUST

TRUST identifies internal repeats in protein sequences by applying transitivity of alignments to detect distant homologous repeats and improve specificity in repeat profiling.

Key Features:

• Transitivity of Alignments: Leverages transitive relationships among alignments to detect repeat homologues that lack direct pairwise alignments.
• Local Suboptimal Alignment Seeding: Begins from significant local suboptimal alignments to extend and link repeat regions via transitivity.
• Noise Reduction and Confidence Enhancement: Aggregates transitive evidence to increase confidence in consistently well-aligned regions and to minimize the impact of non-homologous repeats, producing a generalized repeat profile.
• Rigorous Statistical Testing: Applies statistical tests for both self-sequence and profile-sequence alignments to ensure high specificity in repeat identification.
• Versatile Repeat Prediction: Predicts tandem and non-tandem repeats, including instances with varying intervening segments within a single sequence.
• Virtual Noise-Free Profiling: Constructs profiles intended to represent generalized repeats with reduced noise from spurious matches.

Scientific Applications:

• Functional and Structural Analysis: Identifies internal repeats to aid interpretation of protein structural units and functional domains.
• Evolutionary Studies: Detects repeat duplications and distant homologues to inform studies of protein evolution and duplication events.
• Benchmarking and Validation: Can be evaluated against databases of repeat annotations derived from structural superpositioning for method validation and benchmarking.

Methodology:

Starts from significant local suboptimal alignments, applies transitivity to extend and link alignments into generalized repeat profiles while reducing non-homologous noise, and performs statistical testing for self-sequence and profile-sequence alignments.