همایش بین المللی هوش مصنوعی

صفحه اصلی / اولین همایش بین المللی هوش مصنوعی

Implication of AI programs Alphafold-2 and 3 in Predicting Complex Protein Assemblies: A Case Study on Barnacle Chthamalus malayensis Cement Adhesive Proteins

نویسندگان :

Hosein Moein¹ Maryam Azimzadeh² Aida Arezumand³

1- Faculty of Life Sciences and Biotechnology, Shahid Beheshti University, Tehran, Iran 2- Faculty of Life Sciences and Biotechnology, Shahid Beheshti University Tehran, Iran 3- Faculty Life of Sciences and Biotechnology, Shahid Beheshti University Tehran, Iran

کلمات کلیدی :

AlphaFold،Protein Structure،Multimer،Adhesion،Chthamalus malayensis،Barnacle Cement

چکیده :

High-accuracy protein structure prediction is a prerequisite for understanding biological functions, especially for such complex systems. Herein, we focus on a five-protein adhesive complex from Chthamalus malayensis, a barnacle species that exhibits remarkable adhesion capability. We have predicted the complex's structure and interactions using the AI programs AlphaFold-2 and AlphaFold-3. AlphaFold-3 successfully predicted the complete assembly of the barnacle cement protein complex, accurately integrating all five proteins, including CP20. This provided valuable insights into the molecular interactions within the adhesive complex. In contrast, Alphafold-2 successfully predicted the structures of CP10, CP43, CP52, and CP100 in multimeric structure, but it failed to incorporate CP20 into the multimeric structure, leaving it outside the assembly. This limitation highlights the challenge Alphafold-2 faces in fully assembling complex protein interactions compared to Alphafold-3. Our findings point out the potential of deep-learning-based tools like AlphaFold for the research of marine organisms and allow us to understand how computational models can be further improved to produce more accurate predictions of structural biology.