Moremi Bio Agent: Application of A Foundation Model and End-to-End Automation in the Design and Validation of Monoclonal Antibodies Targeting Plasmodium falciparum Invasion Complex
Akogo, D. A.; Ayensu, J.; Sam, N.; Hattoh, G.; Nyarko, P.; Eshun, S.; Alhasan, M.; Mensah-Brown, H.; Quashie, P.
Show abstract
Malaria remains a significant global health challenge, with Plasmodium falciparum responsible for the majority of severe cases and fatalities. Targeting the parasites invasion mechanisms offers a promising therapeutic strategy. In this study, we leveraged a novel agentic foundational model, Moremi Bio Agent, to design monoclonal antibodies targeting the AMA1-RON2 complex, a critical component in the parasites invasion of human red blood cells. Using advanced structural modeling, we generated 999 antibodies, which were evaluated for binding affinity, structural integrity, and physicochemical properties. Binding affinity analysis using PRODIGY identified 864 antibodies with successful target interactions, exhibiting binding free energies ({Delta}G) ranging from -116.8 kcal/mol to -5.6 kcal/mol. The strongest candidates demonstrated exceptionally tight binding, with dissociation constants (Kd) in the femtomolar to attomolar range, indicative of highly stable interactions. Additionally, structural validation confirmed that the antibodies were thermodynamically stable with robust fold reliability, essential for functional efficacy. Epitope mapping revealed highly conserved regions within the target complex, enhancing the likelihood of cross-strain efficacy. Glycosylation analysis identified key sites that could improve antibody stability and immune recognition, while BLAST comparison with known therapeutic antibodies demonstrated significant homology, underscoring their potential for clinical development. This study highlights the power of generative AI-driven computational pipelines in antibody discovery, providing a scalable and cost-effective framework for therapeutic development. The findings establish a foundation for experimental validation and optimization, with the potential to advance novel interventions against malaria and other infectious diseases.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Computationally Grafting an IgE Epitope onto a Scaffold: Implications for a Pan Anti-Allergy Vaccine Design 94%
- AI-Based Antibody Design Targeting Recent H5N1 Avian Influenza Strains 94%
- DeepSCM: an efficient convolutional neural network surrogate model for the screening of therapeutic antibody viscosity 93%
Similar papers in this journal
- Immuno-informatics Design of a Multimeric Epitope Peptide Based Vaccine Targeting SARS-CoV-2 Spike Glycoprotein 95%
- Predictive Immunoinformatics Reveal Promising Safety and Anti-Onchocerciasis Protective Immune Response Profiles to Vaccine Candidates (Ov-RAL-2 and Ov-103) in Anticipation of Phase I Clinical Trials 94%
- Design of multi-epitope vaccine candidate against Brucella type IV secretion system (T4SS) 94%
Similar papers in this journal
- Utilizing Heteroatom Types and Numbers from Extensive Ligand Libraries to Develop Novel hERG Blocker QSAR Models Using Machine Learning-based Classifiers 92%
- Molecular Glue-Design-Evaluator (MOLDE): An Advanced Method for In-Silico Molecular Glue Design 92%
- Chalcogen derivatives for the treatment of African trypanosomiasis: biological evaluation of thio and seleno- semicarbazones and their azole derivatives 92%
Similar papers in this journal
- Large-scale template-based structural modeling of T-cell receptors with known antigen specificity reveals complementarity features. 92%
- Peptides derived of kunitz-type serine protease inhibitor as potential vaccine against experimental schistosomiasis 92%
- Identification of conserved cross-species B-cell linear epitopes in human malaria: A subtractive proteomics and immuno-informatics approach targeting merozoite stage proteins 91%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.