ApicoAMP: the first computational model for identifying apicoplast-targeted transmembrane proteins in Apicomplexa

Area of Science:

• Computational biology
• Parasitology
• Drug discovery

Background:

• Identifying apicoplast-targeted proteins is crucial for developing drugs against diseases like malaria.
• Existing methods fail to identify apicoplast proteins lacking a bipartite signal, including newly discovered transmembrane proteins.
• This gap necessitates a computational approach to identify this novel class of apicoplast-targeted proteins.

Purpose of the Study:

• To develop a machine learning method for predicting apicoplast-targeted transmembrane proteins in Apicomplexa.
• To create a computational tool that identifies proteins lacking the traditional bipartite signal.

Main Methods:

• Developed an ensemble classification model, ApicoAMP, combining multiple classifiers.
• Utilized feature sets including amino acid hydrophobicity, composition within transmembrane domains, sequence motifs, and Gene Ontology terms.
• Trained the model on proteins from 11 apicomplexan species using a majority vote principle.

Main Results:

• ApicoAMP achieved an overall expected accuracy of 91% in predicting apicoplast-targeted transmembrane proteins.
• The model effectively integrates diverse feature sets for robust classification.
• Demonstrated the feasibility of machine learning for identifying apicoplast proteins with non-canonical targeting signals.

Conclusions:

• ApicoAMP is the first computational model for identifying apicoplast-targeted transmembrane proteins in Apicomplexa.
• The developed software is publicly available for research use.
• This tool advances the computational identification of potential drug targets in parasitic diseases.