Being able to model biomolecular structures is imperative in in silico drug design and hit-to-lead optimization. When crystal protein structures are unavailable, protein structure prediction programs have become a vital tool in scientists’ toolbox. To increase accessibility and function, programs such as AlphaFold31 have arisen to address this need. However, access to many of these tools has remained limited to academia, limiting the use of drug design in commercial spaces. These tools also are limited in predicting small molecule binding motifs. These limitations led to the production of other programs that remove that barrier for scientists, such as RosettaFold All Atom2 from David Baker’s Lab at The University of Washington, Boltz-13 developed out of the CSAIL and Jameel Clinic labs at MIT and Chai-14 from the Chai Discovery Team.

BMaps Integration and Use Cases For AI Protein Folding
Our newest release of Boltzmann Maps (BMaps) now includes integration of two new open-source AI protein folding programs, Boltz-1 and Chai-1. Both protein-sequence specific models rival the predictions generated by programs such as AlphaFold3 (Figure 1). We integrate these programs in our “Choose Protein” window. From there, there is a “Protein Folding” tab where there are a variety of options to control in the users’ workflow. (Figure 2). Benchmarking internally has shown success in a variety of use cases incorporating ligands and ions. The alignment scores perform very well against well-annotated PDB structures. Additionally, since the simulations have similar run times to that of docking methods they can be used as an orthogonal method for docking pose generation.

Model-Ready, Enriched Protein Prediction


In addition to small molecules and ions, BMaps can include cofactors, lipids, and even water molecules when folding with Boltz-1 and Chai-1. Figure 3 and Figure 4 include examples of protein folding completed in BMaps that accurately predict the location of a Zn2+ ion. These structural ions and other cofactors assist in predicting more biologically relevant structures.
With AI protein folding, atoms are mapped to coordinates without secondary structure information and does not store bond order information or implicit hydrogens so the default output from the folding programs require significant preparation before the structures are ready for modeling. BMaps however integrates secondary structure assignments and properly assigns valences and adds hydrogens, making the structures ready for modeling tasks. This is all done automatically on our servers, so the user does not have to go through these difficult tasks. Correct protein and ligand structure predictions allow users to maintain a seamless workflow all in BMaps.
Structure-based design has a new tool in the toolbox for scientists of all levels, easily integrated into BMaps. Use cases for Boltz-1 or Chai-1 can both: lead to structure predictions in which there are no experimental structures, or for alternate poses for a ligand where there are known structures. The utility of these programs unified within BMaps allows for users to have a one-stop-shop for drug design.
References
[1] Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O., Willmore, L., Ballard, A. J., Bambrick, J., Bodenstein, S. W., Evans, D. A., Hung, C.-C., O’Neill, M., Reiman, D., Tunyasuvunakool, K., Wu, Z., Žemgulytė, A., Arvaniti, E., … Jumper, J. M. (2024). Accurate structure prediction of biomolecular interactions with alphafold 3. Nature, 630(8016), 493–500. https://doi.org/10.1038/s41586-024-07487-w
[2] Krishna, R., Wang, J., Ahern, W., Sturmfels, P., Venkatesh, P., Kalvet, I., Lee, G. R., Morey-Burrows, F. S., Anishchenko, I., Humphreys, I. R., McHugh, R., Vafeados, D., Li, X., Sutherland, G. A., Hitchcock, A., Hunter, C. N., Kang, A., Brackenbrough, E., Bera, A. K., … Baker, D. (2024). Generalized biomolecular modeling and design with Rosettafold All-Atom. Science, 384(6693). https://doi.org/10.1126/science.adl2528
[3] Wohlwend, J., Corso, G., Passaro, S., Reveiz, M., Leidal, K., Swiderski, W., Portnoi, T., Chinn, I., Silterra, J., Jaakkola, T., & Barzilay, R. (2024). Boltz-1: Democratizing biomolecular interaction modeling. bioRxiv. https://doi.org/10.1101/2024.11.19.624167
[4] Boitreaud, J., Dent, J., McPartlon, M., Meier, J., Reis, V., Rogozhnikov, A., & Wu, K. (2024). Chai-1: Decoding the molecular interactions of life. bioRxiv. https://doi.org/10.1101/2024.10.10.615955