ConSurf-DB

ConSurf-DB provides evolutionary conservation profiles for proteins with known structures from the Protein Data Bank (PDB) to identify functionally and structurally important amino acid positions.

Key Features:

• PDB-wide precalculated profiles: Precalculated conservation analyses are available for nearly all structures deposited in the PDB.
• Input types: Analyses can be initiated from either a protein sequence or a 3D structure.
• Homolog search and MSA generation: The pipeline searches for similar sequences and aligns them into a multiple sequence alignment (MSA).
• Per-site evolutionary rates: Evolutionary rates are calculated at each amino acid site to produce conservation profiles.
• Bayesian and maximum-likelihood algorithms: Rate estimation uses Bayesian or maximum-likelihood methods that incorporate phylogenetic trees.
• Sampling correction: The algorithms address uneven sampling in sequence space to mitigate biases from limited sequence data.
• Structure mapping: Conservation scores are mapped onto protein structures to highlight conserved residues in 3D context.

Scientific Applications:

• Functional residue identification: Pinpoints amino acid positions likely to be critical for protein function and stability.
• Mutation interpretation: Assists in assessing the potential functional impact of individual mutations.
• Binding and catalytic site annotation: Supports identification of binding, catalytic, and transport-related residues on structures.
• Large-scale analyses: Enables genome-wide or large-scale investigations of protein conservation patterns across many structures.
• Evolutionary and structural studies: Facilitates studies of protein evolution, native conformation maintenance, and biological activity conservation.

Methodology:

Given a query sequence or 3D structure, ConSurf-DB searches for similar sequences, constructs a multiple sequence alignment (MSA), and computes per-site evolutionary rates using Bayesian or maximum-likelihood algorithms that incorporate phylogenetic trees and account for uneven sampling in sequence space.