CroMaSt: a workflow for assessing protein domain classification by cross-mapping of structural instances between...

Area of Science:

• Structural Biology
• Bioinformatics
• Computational Biology

Background:

• Protein domains are fundamental units for understanding protein structure-function relationships.
• Discrepancies in domain definitions and boundaries across databases hinder accurate protein domain classification.
• This inconsistency raises challenges in defining and enumerating true protein domain instances.

Purpose of the Study:

• To develop an automated workflow for assessing and standardizing protein domain classification across different databases.
• To address the issue of varying domain definitions and boundaries by cross-mapping structural instances.
• To provide a robust method for classifying experimental structural instances of protein domains.

Main Methods:

• An automated iterative workflow named CroMaSt (Cross-Mapper of domain Structural instances) was developed.
• CroMaSt cross-maps domain structural instances between major databases (Pfam and CATH) using structural alignments.
• The workflow utilizes the Kpax structural alignment tool with expert-adjusted parameters and classifies instances into 'Core', 'True', 'Domain-like', and 'Failed' categories.

Main Results:

• CroMaSt was successfully tested on the RNA Recognition Motif domain type.
• The workflow identified 962 'True' and 541 'Domain-like' structural instances for this domain.
• This method offers a standardized approach to domain classification, resolving critical issues in domain-centric research.

Conclusions:

• The developed CroMaSt workflow provides a reliable method for assessing protein domain classification.
• This approach can generate crucial data for synthetic biology and machine learning-driven protein domain engineering.
• Standardized domain classification facilitates more accurate structure-function relationship studies and protein design.