ProtPen combines sequence- and structure-based approaches to facilitate protein function predictions on a...

Area of Science:

• Proteomics and Bioinformatics
• Molecular Biology

Background:

• Proteins of unknown function are a major knowledge gap, especially in prokaryotes.
• Understanding these proteins is crucial for biological and pathogenicity studies.

Purpose of the Study:

• To introduce ProtPen, an open-source pipeline for predicting protein function.
• To integrate sequence-based and structure-based methods for comprehensive annotation.

Main Methods:

• ProtPen combines eggNOG-mapper (sequence-based) and Foldseek (structure-based) using AlphaFold structures.
• It merges results and enriches with UniProt metadata for detailed analysis.
• The pipeline accepts FASTA files with UniProt identifiers and scales to whole proteomes.

Main Results:

• Benchmarking on *Pseudomonas aeruginosa* showed >90% annotation accuracy.
• Demonstrated complementarity of sequence- and structure-based prediction methods.
• Successfully applied to identify functions in differentially abundant proteins and uncharacterized glycoproteins.

Conclusions:

• ProtPen facilitates system-wide annotation of proteins with unknown functions.
• It is a valuable tool for analyzing proteomic datasets and whole proteomes.
• The pipeline is extensible for incorporating additional prediction tools.