Searching for protein signatures using a multilevel alphabet

Area of Science:

• Biochemistry
• Structural Biology
• Bioinformatics

Background:

• Short motifs in proteins are crucial for diverse biological functions, including molecular interactions and membrane binding.
• Current methods for motif detection primarily rely on sequence information, overlooking structural and physicochemical properties.

Purpose of the Study:

• To develop a novel approach for identifying protein signatures that integrate sequence, structural, and physicochemical properties.
• To detect enriched structural motifs and predict protein properties from sequence data.
• To identify novel interacting motifs within the yeast protein interactome.

Main Methods:

• Translating amino acid sequences into a new alphabet reflecting intrinsic structural and chemical properties.
• Utilizing the MEME search algorithm to identify protein signatures within subsets of proteins with common sequence and structural information.
• Applying the method to the yeast protein interactome dataset.

Main Results:

• Successfully detected enriched structural motifs, such as amphipathic helices, from large datasets of linear sequences.
• Accurately predicted common structural properties (disorder, surface accessibility, secondary structures) of known functional motifs.
• Identified novel putative interacting motifs within the yeast protein interactome.

Conclusions:

• The developed method effectively identifies protein signatures by considering diverse properties beyond sequence alone.
• This approach holds potential for de novo protein function prediction using sequence or structural information.
• The findings contribute to a deeper understanding of protein motif function and interactions.