Sequence-structure-function relationships in the microbial protein universe

Area of Science:

• Structural Biology
• Bioinformatics
• Genomics

Background:

• Traditional structural biology assumes sequence similarity dictates structure and function.
• This assumption limits exploration of the full protein universe.
• Alternative pathways for protein function exist, independent of sequence homology.

Purpose of the Study:

• To explore protein universe regions where function is independent of sequence and structure.
• To predict and functionally annotate protein structures for diverse microbial proteomes.
• To create a database complementing existing resources like AlphaFold.

Main Methods:

• Prediction of ~200,000 protein structures using the World Community Grid.
• Functional annotation of predicted structures on a per-residue basis.
• Comparative analysis with the AlphaFold database for coverage and diversity.

Main Results:

• Identification of 148 novel protein folds.
• Mapping of specific functions to structural motifs.
• Demonstration of a continuous and saturated protein structural space.
• Creation of a database with broad genomic and sequence diversity.

Conclusions:

• Protein function can be achieved through diverse sequences and structures.
• A shift towards sequence-structure-function based meta-omics is needed.
• The predicted structural models offer complementary insights to existing databases.