Sequence statistics of tertiary structural motifs reflect protein stability

Area of Science:

• Computational structural biology
• Protein bioinformatics
• Biophysics

Background:

• The Protein Data Bank (PDB) is crucial for understanding protein sequence-structure relationships.
• Traditional methods use geometric descriptors to analyze these relationships.
• The PDB's continuous growth presents opportunities for novel analytical approaches.

Purpose of the Study:

• To explore a new method for deducing quantitative sequence-structure relationships by analyzing complex structural motifs.
• To validate this approach by predicting protein stability changes upon amino acid substitution (ΔΔGm).
• To assess the performance of this method against existing state-of-the-art techniques.

Main Methods:

• Defined non-contiguous tertiary motifs (TERMs) around specific protein sites.
• Extracted sequence preferences from ensembles of closely matching substructures in the PDB.
• Used these ensemble statistics to predict mutational stability changes (ΔΔGm).

Main Results:

• The TERM-based method predicts ΔΔGm comparably to leading statistical and machine-learning methods on large datasets.
• It outperforms existing methods on diverse, unbiased mutations.
• Performance directly correlates with the amount of relevant structural data in the PDB, improving as the PDB grows.

Conclusions:

• Statistics of non-contiguous structural motifs in the PDB encode fundamental sequence-structure relationships relevant to protein thermodynamic stability.
• The PDB is sufficiently large for these statistics to be practically useful, with accuracy expected to increase.
• This approach offers new avenues for utilizing structural data in computational structural biology.