MoBiTDE: A Molecular Bilinear Two-Tower Network for Drug-Excipient Nanoparticle Co-Assembly Prediction
DOI:
https://doi.org/10.62051/spm3t666Keywords:
co-assembly; SMI-TED; MoBiTDE; Chemical Large Language Model; Insoluble Drugs.Abstract
Up to 40% of clinical candidates are poorly-soluble drugs, which rely heavily on nanoparticle formulations for oral and targeted delivery. Successful formulation largely depends on the co-assembly of drug and excipients into stable nanostructures. However, traditional wet-lab trial-and-error approaches require months, even years for discoveries to be made, with a substantial cost in raw materials. From research, there are 1440 high-throughput screening pairs of drug combination, but few AI algorithms have been applied to this task, especially with the rapidly evolving chemistry large language model (LLM), which hasn’t been evaluated on this task yet. Herein, I introduce the novel application of chemistry foundation model representations to drug-excipient nanoparticle co-assembly prediction. Specifically, this essay introduces MoBiTDE, a molecular bilinear two-tower network designed to learn pairwise compatibility between drugs and excipients. MoBiTDE employs frozen SMI-TED embeddings as pre-trained molecular representations, complemented by 24 RDKit-derived physicochemical descriptors. Drugs and excipient representations are processed through two independent molecular towers and coupled by a rank-32 bilinear interaction module to explicitly capture cross-molecular interactions. The resulting pairwise representation is subsequently fed into a single positive-weighted binary classification head for co-assembly prediction. Experimental results demonstrate that MoBiTDE outperforms other methods under different evaluation settings. Notably, its predictive performance further improves as the dataset expands, highlighting the potential of pre-trained molecular representations to benefit from increasing experimental data. In summary, MoBiTDE provides an effective framework for learning drug–excipient compatibility and offers a promising computational approach to facilitate the discovery of self-assembling nanoparticle formulations. It can improve drug discovery efficiency by reducing the number of trials, focusing only wet-lab testing on promising combinations.
Downloads
References
[1] G. Gadaleta-Caldarola et al., “Sorafenib: 10 years after the first pivotal trial,” Future Oncol., vol. 11, no. 13, pp. 1863–1880, Jul. 2015, doi: 10.2217/fon.15.85. DOI: https://doi.org/10.2217/fon.15.85
[2] M. Makanga and S. Krudsood, “The clinical efficacy of artemether/lumefantrine (Coartem®),” Malar. J., vol. 8, no. S1, p. S5, Dec. 2009, doi: 10.1186/1475-2875-8-S1-S5. DOI: https://doi.org/10.1186/1475-2875-8-S1-S5
[3] A. Deißler et al., “Rapamycin delays allograft rejection in obese graft recipients through induction of myeloid-derived suppressor cells,” Immunol. Lett., vol. 236, pp. 1–11, Aug. 2021, doi: 10.1016/j.imlet.2021.05.003. DOI: https://doi.org/10.1016/j.imlet.2021.05.003
[4] S. Kalepu and V. Nekkanti, “Insoluble drug delivery strategies: review of recent advances and business prospects,” Acta Pharm. Sin. B, vol. 5, no. 5, pp. 442–453, Sep. 2015, doi: 10.1016/j.apsb.2015.07.003. DOI: https://doi.org/10.1016/j.apsb.2015.07.003
[5] G. L. Amidon, H. Lennernäs, V. P. Shah, and J. R. Crison, “A Theoretical Basis for a Biopharmaceutic Drug Classification: The Correlation of in Vitro Drug Product Dissolution and in Vivo Bioavailability,” Pharm. Res., vol. 12, no. 3, pp. 413–420, Mar. 1995, doi: 10.1023/A:1016212804288. DOI: https://doi.org/10.1023/A:1016212804288
[6] A. R. Hashmi, M. Sekar, F. Zahra, N. Molugulu, and L. S. Wong, “Advanced Drug Delivery Strategies to Overcome Solubility and Permeability Challenges: Driving Biopharmaceutical Advancements Toward Commercial Success,” ACS Omega, vol. 10, no. 36, pp. 40769–40792, Sep. 2025, doi: 10.1021/acsomega.5c06171. DOI: https://doi.org/10.1021/acsomega.5c06171
[7] G. M. Whitesides and B. Grzybowski, “Self-Assembly at All Scales,” Science, vol. 295, no. 5564, pp. 2418–2421, Mar. 2002, doi: 10.1126/science.1070821. DOI: https://doi.org/10.1126/science.1070821
[8] O. C. Farokhzad and R. Langer, “Impact of Nanotechnology on Drug Delivery,” ACS Nano, vol. 3, no. 1, pp. 16–20, Jan. 2009, doi: 10.1021/nn900002m. DOI: https://doi.org/10.1021/nn900002m
[9] Y. Xu, J. Pei, and L. Lai, “Deep Learning Based Regression and Multiclass Models for Acute Oral Toxicity Prediction with Automatic Chemical Feature Extraction,” J. Chem. Inf. Model., vol. 57, no. 11, pp. 2672–2685, Nov. 2017, doi: 10.1021/acs.jcim.7b00244. DOI: https://doi.org/10.1021/acs.jcim.7b00244
[10] H. Öztürk, A. Özgür, and E. Ozkirimli, “DeepDTA: deep drug–target binding affinity prediction,” Bioinformatics, vol. 34, no. 17, pp. i821–i829, Sep. 2018, doi: 10.1093/bioinformatics/bty593.
[11] D. Reker et al., “Computationally guided high-throughput design of self-assembling drug nanoparticles,” Nat. Nanotechnol., vol. 16, no. 6, pp. 725–733, Jun. 2021, doi: 10.1038/s41565-021-00870-y. DOI: https://doi.org/10.1038/s41565-021-00870-y
[12] K. Huang, C. Xiao, L. M. Glass, and J. Sun, “MolTrans: Molecular Interaction Transformer for drug–target interaction prediction,” Bioinformatics, vol. 37, no. 6, pp. 830–836, May 2021, doi: 10.1093/bioinformatics/btaa880. DOI: https://doi.org/10.1093/bioinformatics/btaa880
[13] J.-H. Kim, J. Jun, and B.-T. Zhang, “Bilinear Attention Networks,” Oct. 19, 2018, arXiv: arXiv:1805.07932. doi: 10.48550/arXiv.1805.07932.
[14] J. Ross, B. Belgodere, V. Chenthamarakshan, I. Padhi, Y. Mroueh, and P. Das, “Large-scale chemical language representations capture molecular structure and properties,” Nat. Mach. Intell., vol. 4, no. 12, pp. 1256–1264, Dec. 2022, doi: 10.1038/s42256-022-00580-7. DOI: https://doi.org/10.1038/s42256-022-00580-7
[15] Y. Xiang et al., “Physics-informed design of drug-excipient nanoparticles via free energy calculation and yoked deep learning,” Npj Comput. Mater., Aug. 2026, doi: 10.1038/s41524-026-02262-z. DOI: https://doi.org/10.1038/s41524-026-02262-z
[16] E. Soares, E. Vital Brazil, V. Shirasuna, D. Zubarev, R. Cerqueira, and K. Schmidt, “An open-source family of large encoder-decoder foundation models for chemistry,” Commun. Chem., vol. 8, no. 1, p. 193, Jul. 2025, doi: 10.1038/s42004-025-01585-0. DOI: https://doi.org/10.1038/s42004-025-01585-0
[17] T.-Y. Lin, A. RoyChowdhury, and S. Maji, “Bilinear CNN Models for Fine-Grained Visual Recognition,” in 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, Chile: IEEE, Dec. 2015, pp. 1449–1457. doi: 10.1109/ICCV.2015.170. DOI: https://doi.org/10.1109/ICCV.2015.170
[18] H. Öztürk, A. Özgür, and E. Ozkirimli, “DeepDTA: deep drug–target binding affinity prediction,” Bioinformatics, vol. 34, no. 17, pp. i821–i829, Sep. 2018, doi: 10.1093/bioinformatics/bty593. DOI: https://doi.org/10.1093/bioinformatics/bty593
[19] J.-H. Kim, J. Jun, and B.-T. Zhang, “Bilinear Attention Networks,” Oct. 19, 2018, arXiv: arXiv:1805.07932. doi: 10.48550/arXiv.1805.07932.
[20] J. M. Johnson and T. M. Khoshgoftaar, “Survey on deep learning with class imbalance,” J. Big Data, vol. 6, no. 1, p. 27, Dec. 2019, doi: 10.1186/s40537-019-0192-5. DOI: https://doi.org/10.1186/s40537-019-0192-5
[21] S. Chithrananda, G. Grand, and B. Ramsundar, “ChemBERTa: Large-Scale Self-Supervised Pretraining for Molecular Property Prediction,” Oct. 23, 2020, arXiv: arXiv:2010.09885. doi: 10.48550/arXiv.2010.09885.
[22] E. J. Hu et al., “LoRA: Low-Rank Adaptation of Large Language Models,” Oct. 16, 2021, arXiv: arXiv:2106.09685. doi: 10.48550/arXiv.2106.09685.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Transactions on Materials, Biotechnology and Life Sciences
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
