Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Directional multiobjective optimization of metal complexes at the billion-system scale

Nature Computational Science volume 4, pages 263–273 (2024)Cite this article

Subjects

A preprint version of the article is available at ChemRxiv.

Abstract

The discovery of transition metal complexes (TMCs) with optimal properties requires large ligand libraries and efficient multiobjective optimization algorithms. Here we provide the tmQMg-L library, containing 30k diverse and synthesizable ligands with robustly assigned charges and metal coordination modes. tmQMg-L enabled the generation of 1.37 million palladium TMCs, which were used to develop and benchmark the Pareto-Lighthouse multiobjective genetic algorithm (PL-MOGA). With fine control over aim and scope, this algorithm maximized both the polarizability and highest occupied molecular orbital–lowest unoccupied molecular orbital gap of the TMCs within selected regions of the Pareto front, without requiring prior knowledge on the objective limits. Instead of genetic operations on small ligand fragments, the PL-MOGA did whole-ligand mutation and crossover operations, which in chemical spaces containing billions of systems, yielded thousands of highly diverse TMCs in an interpretable manner.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: PL-MOGA algorithm.
Fig. 2: Derivation of the tmQMg-L ligand dataset.
Fig. 3: Size and distribution of the chemical spaces.
Fig. 4: Multiobjective (α, ϵ) optimizations within the 1.37M space.
Fig. 5: Interpretability and diversity.
Fig. 6: Pareto front distributions and samples in the billion spaces.

Similar content being viewed by others

Data availability

The tmQMg-L dataset can be accessed at https://github.com/hkneiding/tmQMg-L, including the data for the charge assignment benchmark and the 1.37M space. The dataset is also available via Zenodo at https://doi.org/10.5281/zenodo.10374523 (ref. 64). In addition to the geometric and electronic structure information, it provides Weisfeiler–Lehman graph hashes65. All data are openly available. Source data are provided with this paper. The larger datasets may require Linux software to be visualized.

Code availability

The PL-MOGA code is available from https://github.com/hkneiding/PL-MOGA and Zenodo via https://doi.org/10.5281/zenodo.10663863 (ref. 66), including the DFT geometries of selected TMC hits and the weighted-sum benchmark. The code includes a command line functionality, together with documentation and installation instructions. All code is openly available.

References

  1. Mjos, K. D. & Orvig, C. Metallodrugs in medicinal inorganic chemistry. Chem. Rev. 114, 4540–4563 (2014).

    Google Scholar 

  2. Prier, C. K., Rankic, D. A. & MacMillan, D. W. C. Visible light photoredox catalysis with transition metal complexes: applications in organic synthesis. Chem. Rev. 113, 5322–5363 (2013).

    Google Scholar 

  3. Kalyanasundaram, K. & Gratzel, M. Applications of functionalized transition metal complexes in photonic and optoelectronic devices. Coord. Chem. Rev. 177, 347–414 (1998).

    Google Scholar 

  4. Yoon, T. P., Ischay, M. A. & Du, J. N. Visible light photocatalysis as a greener approach to photochemical synthesis. Nature Chem. 2, 527–532 (2010).

    Google Scholar 

  5. Furukawa, H., Cordova, K. E., O’Keeffe, M. & Yaghi, O. M. The chemistry and applications of metal–organic frameworks. Science 341, 974 (2013).

    Google Scholar 

  6. Balcells, D. & Nova, A. Designing Pd and Ni catalysts for cross-coupling reactions by minimizing off-cycle species. ACS Catal. 8, 3499–3515 (2018).

    Google Scholar 

  7. Foscato, M. & Jensen, V. R. Automated in silico design of homogeneous catalysts. ACS Catal. 10, 2354–2377 (2020).

    Google Scholar 

  8. Robbins, D. W. & Hartwig, J. F. A simple, multidimensional approach to high-throughput discovery of catalytic reactions. Science 333, 1423–1427 (2011).

    Google Scholar 

  9. Nandy, A. et al. Computational discovery of transition-metal complexes: from high-throughput screening to machine learning. Chem. Rev. 121, 9927–10000 (2021).

    Google Scholar 

  10. Huang, B. & von Lilienfeld, O. A. Ab initio machine learning in chemical compound space. Chem. Rev. 121, 10001–10036 (2021).

    Google Scholar 

  11. Freeze, J. G., Kelly, H. R. & Batista, V. S. Search for catalysts by inverse design: artificial intelligence, mountain climbers, and alchemists. Chem. Rev. 119, 6595–6612 (2019).

    Google Scholar 

  12. Kitchin, J. R. Machine learning in catalysis. Nat. Catal. 1, 230–232 (2018).

    Google Scholar 

  13. Gomes, G. D., Pollice, R. & Aspuru-Guzik, A. Navigating through the maze of homogeneous catalyst design with machine learning. Trends Chem. 3, 96–110 (2021).

    Google Scholar 

  14. Friederich, P., Gomes, G. D., De Bin, R., Aspuru-Guzik, A. & Balcells, D. Machine learning dihydrogen activation in the chemical space surrounding Vaska’s complex. Chem. Sci. 11, 4584–4601 (2020).

    Google Scholar 

  15. Nandy, A., Duan, C. R., Goffinet, C. & Kulik, H. J. New strategies for direct methane-to-methanol conversion from active learning exploration of 16 million catalysts. JACS Au 2, 1200–1213 (2022).

    Google Scholar 

  16. Jorner, K., Tomberg, A., Bauer, C., Skold, C. & Norrby, P. O. Organic reactivity from mechanism to machine learning. Nat. Rev. Chem. 5, 240–255 (2021).

    Google Scholar 

  17. Vamathevan, J. et al. Applications of machine learning in drug discovery and development. Nat. Rev. Drug Discov. 18, 463–477 (2019).

    Google Scholar 

  18. Butler, K. T., Davies, D. W., Cartwright, H., Isayev, O. & Walsh, A. Machine learning for molecular and materials science. Nature 559, 547–555 (2018).

    Google Scholar 

  19. Goldberg, D. E. Genetic Algorithms in Search, Optimization, and Machine Learning (Addison-Wesley, 1989).

  20. De Jong, K. A. Evolutionary Computation—A Unified Appraoch (MIT Press, 2006).

  21. Winter, R. et al. Efficient multi-objective molecular optimization in a continuous latent space. Chem. Sci. 10, 8016–8024 (2019).

    Google Scholar 

  22. Anstine, D. M. & Isayev, O. Generative models as an emerging paradigm in the chemical sciences. J. Am. Chem. Soc. 145, 8736–8750 (2023).

    Google Scholar 

  23. Sanchez-Lengeling, B. & Aspuru-Guzik, A. Inverse molecular design using machine learning: generative models for matter engineering. Science 361, 360–365 (2018).

    Google Scholar 

  24. Le, T. C. & Winkler, D. A. Discovery and optimization of materials using evolutionary approaches. Chem. Rev. 116, 6107–6132 (2016).

    Google Scholar 

  25. Jensen, J. H. A graph-based genetic algorithm and generative model/Monte Carlo tree search for the exploration of chemical space. Chem. Sci. 10, 3567–3572 (2019).

    Google Scholar 

  26. Nigam, A., Pollice, A. & Aspuru-Guzik, A. Parallel tempered genetic algorithm guided by deep neural networks for inverse molecular design. Digit. Discov. 1, 390–404 (2022).

    Google Scholar 

  27. Janet, J. P., Chan, L. & Kulik, H. J. Accelerating chemical discovery with machine learning: simulated evolution of spin crossover complexes with an artificial neural network. J. Phys. Chem. Lett. 9, 1064–1071 (2018).

    Google Scholar 

  28. Gallarati, S., Gerwen, P. V., Schoepfer, A. A., Laplaza, R. & Corminboeuf, C. Genetic algorithms for the discovery of homogeneous catalysts. CHIMIA 77, 39 (2023).

    Google Scholar 

  29. Fey, N., Orpen, A. G. & Harvey, J. N. Building ligand knowledge bases for organometallic chemistry: computational description of phosphorus(III)-donor ligands and the metal-phosphorus bond. Coord. Chem. Rev. 253, 704–722 (2009).

    Google Scholar 

  30. Gugler, S., Janet, J. P. & Kulik, H. J. Enumeration of de novo inorganic complexes for chemical discovery and machine learning. Mol. Syst. Des. Eng. 5, 139–152 (2020).

    Google Scholar 

  31. Gensch, T. et al. A comprehensive discovery platform for organophosphorus ligands for catalysis. J. Am. Chem. Soc. 144, 1205–1217 (2022).

    Google Scholar 

  32. Ioannidis, E. I., Gani, T. Z. H. & Kulik, H. J. molSimplify: a toolkit for automating discovery in inorganic chemistry. J. Comput. Chem. 37, 2106–2117 (2016).

    Google Scholar 

  33. Foscato, M., Venkatraman, V. & Jensen, V. R. DENOPTIM: software for computational de novo design of organic and inorganic molecules. J. Chem. Inf. Model. 59, 4077–4082 (2019).

    Google Scholar 

  34. Sobez, J. G. & Reiher, M. MOLASSEMBLER: molecular graph construction, modification, and conformer generation for inorganic and organic molecules. J. Chem. Inf. Model. 60, 3884–3900 (2020).

    Google Scholar 

  35. Chen, S. et al. Automated construction and optimization combined with machine learning to generate Pt(II) methane C–H activation transition states. Top. Catal. 65, 312–324 (2022).

    Google Scholar 

  36. Kneiding, H. et al. Deep learning metal complex properties with natural quantum graphs. Digit. Discov. 2, 618–633 (2023).

    Google Scholar 

  37. Groom, C. R., Bruno, I. J., Lightfoot, M. P. & Ward, S. C. The Cambridge Structural Database. Acta Cryst. B B72, 171–179 (2016).

    Google Scholar 

  38. Duan, C. et al. Exploiting ligand additivity for transferable machine learning of multireference character across known transition metal complex ligands. J. Chem. Theory Comput. 18, 4836–4845 (2022).

    Google Scholar 

  39. Vela, S., Laplaza, R., Cho, Y. R. & Corminboeuf, C. cell2mol: encoding chemistry to interpret crystallographic data. Npj Comput. Mater. 8, 188 (2022).

    Google Scholar 

  40. Matsuoka, W., Harabuchi, Y. & Maeda, S. Virtual ligand-assisted screening strategy to discover enabling ligands for transition metal catalysis. ACS Catal. 12, 3752–3766 (2022).

    Google Scholar 

  41. Gao, W. H. & Coley, C. W. The synthesizability of molecules proposed by generative models. J. Chem. Inf. Model. 60, 5714–5723 (2020).

    Google Scholar 

  42. Chu, Y. H., Heyndrickx, W., Occhipinti, G., Jensen, V. R. & Alsberg, B. K. An evolutionary algorithm for de novo optimization of functional transition metal compounds. J. Am. Chem. Soc. 134, 8885–8895 (2012).

    Google Scholar 

  43. Durrant, M. C. The use of quantum molecular calculations to guide a genetic algorithm: a way to search for new chemistry. Chem. Eur. J. 13, 3406–3413 (2007).

    Google Scholar 

  44. Janet, J. P., Ramesh, S., Duan, C. & Kulik, H. J. Accurate multiobjective design in a space of millions of transition metal complexes with neural-network-driven efficient global optimization. ACS Cent. Sci. 6, 513–524 (2020).

    Google Scholar 

  45. Sowndarya, S. V. S. et al. Multi-objective goal-directed optimization of de novo stable organic radicals for aqueous redox flow batteries. Nat. Mach. Intell. 4, 720–730 (2022).

    Google Scholar 

  46. Verhellen, J. Graph-based molecular Pareto optimisation. Chem. Sci. 13, 7526–7535 (2022).

    Google Scholar 

  47. Hase, F., Roch, L. M. & Aspuru-Guzik, A. Chimera: enabling hierarchy based multi-objective optimization for self-driving laboratories. Chem. Sci. 9, 7642–7655 (2018).

    Google Scholar 

  48. Nigam, A., Pollice, R., Krenn, M., Gomes, G. D. & Aspuru-Guzik, A. Beyond generative models: superfast traversal, optimization, novelty, exploration and discovery (STONED) algorithm for molecules using SELFIES. Chem. Sci. 12, 7079–7090 (2021).

    Google Scholar 

  49. Brown, N., Fiscato, M., Segler, M. H. S. & Vaucher, A. C. GuacaMol: benchmarking models for de novo molecular design. J. Chem. Inf. Model. 59, 1096–1108 (2019).

    Google Scholar 

  50. Laplaza, R., Gallarati, S. & Corminboeuf, C. Genetic optimization of homogeneous catalysts. Chem. Methods 2, e202100107 (2022).

    Google Scholar 

  51. Seumer, J., Hansen, J. K. S., Nielsen, M. B. & Jensen, J. H. Computational evolution of new catalysts for the Morita–Baylis–Hillman reaction. Angew. Chem. Int. Ed. 62, e202218565 (2023).

    Google Scholar 

  52. Balcells, D. & Skjelstad, B. B. tmQM dataset–quantum geometries and properties of 86k transition metal complexes. J. Chem. Inf. Model. 60, 6135–6146 (2020).

    Google Scholar 

  53. Chen, S. et al. ReaLigands: a ligand library cultivated from experiment and intended for molecular computational catalyst design. J. Chem. Inf. Model. https://doi.org/10.1021/acs.jcim.3c01310 (2023).

  54. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).

    Google Scholar 

  55. Weigend, F. & Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: design and assessment of accuracy. Phys. Chem. Chem. Phys. 7, 3297–3305 (2005).

    Google Scholar 

  56. von Lilienfeld, O. A., Müller, K. R. & Tkatchenko, A. Exploring chemical compound space with quantum-based machine learning. Nat. Rev. Chem. 4, 347–358 (2020).

    Google Scholar 

  57. Hoffmeister, F. & Sprave, J. Problem-independent handling of constraints by use of metric penalty functions. In Evolutionary Programing (1996); https://ls11-www.cs.tu-dortmund.de/~joe/papers/ep96a.pdf

  58. Devi, R. V., Sathya, S. S. & Coumar, M. S. Multi-objective genetic algorithm for de novo drug design (MoGADdrug). Curr. Comput. Aid. Drug Des. 17, 445–457 (2021).

    Google Scholar 

  59. Pollice, R. et al. Data-driven strategies for accelerated materials design. Acc. Chem. Res. 54, 849–860 (2021).

    Google Scholar 

  60. Hueffel, J. A. et al. Accelerated dinuclear palladium catalyst identification through unsupervised machine learning. Science 374, 1134–1140 (2021).

    Google Scholar 

  61. Adamo, A. & Barone, V. Toward reliable density functional methods without adjustable parameters: the PBE0 model. J. Chem. Phys. 110, 6158–6169 (1999).

    Google Scholar 

  62. Grimme, S., Bannwarth, C. & Shushkov, P. A robust and accurate tight-binding quantum chemical method for structures, vibrational frequencies, and noncovalent interactions of large molecular systems parametrized for all spd-block elements (Z = 1–86). J. Chem. Theory Comput. 13, 1989–2009 (2017).

    Google Scholar 

  63. Bannwarth, C., Ehlert, S. & Grimme, S. GFN2-xTB—an accurate and broadly parametrized self-consistent tight-binding quantum chemical method with multipole electrostatics and density-dependent dispersion contributions. J. Chem. Theory Comput. 15, 1652–1671 (2019).

    Google Scholar 

  64. Kneiding, H., Balcells, D. & Nova, A. tmQMg-L. Zenodo https://doi.org/10.5281/zenodo.10374523 (2023).

  65. Nandy, A., Taylor, M. G. & Kulik, H. J. Identifying underexplored and untapped regions in the chemical space of transition metal complexes. J. Phys. Chem. Lett. 14, 5798–5804 (2023).

    Google Scholar 

  66. Kneiding, H. tmQMg-L. Zenodo https://doi.org/10.5281/zenodo.10663863 (2024).

Download references

Acknowledgements

European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement number 945371 (H.K.). This article reflects only the author’s view and the REA is not responsible for any use that may be made of the information it contains. Research Council of Norway (RCN) FRIPRO program supporting the CO2pCat project, with number 314321 (A.N.). RCN FRIPRO program supporting the catLEGOS project, with number 325003 (D.B.). RCN support through the Centers of Excellence program, including the Hylleraas Centre, with project number 262695, and the Sigma2 – National Infrastructure for High Performance Computing and Data Storage in Norway, with grant number NN4654K (H.K., A.N. and D.B.). We also thank M. Strandgaard and T. Linjordet for helpful discussions and for reviewing preliminary versions of this manuscript.

Author information

Authors and Affiliations

  1. Hylleraas Centre for Quantum Molecular Sciences, Department of Chemistry, University of Oslo, Oslo, Norway

    Hannes Kneiding, Ainara Nova & David Balcells

  2. Centre for Materials Science and Nanotechnology, Department of Chemistry, University of Oslo, Oslo, Norway

    Ainara Nova

Authors
  1. Hannes Kneiding
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Ainara Nova
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. David Balcells
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

H.K. was the main developer of the tmQMg-L dataset, the 1.37M space and the PL-MOGA algorithm. H.K. also derived the combinatorics of the square planar TMC space and developed the concept of a generative model based on whole-ligand multiple-site genetic operations. A.N. and D.B. developed the concept of extracting the ligand charges from the natural Lewis structures. All authors made substantial contributions to the conception and design of the work. D.B. was the main contributor to the writing and revision of the manuscript, as well as to the definition, supervision and funding of the research project.

Corresponding author

Correspondence to David Balcells.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Computational Science thanks Jan Jensen, Aditya Nandy and Robert Pollice for their contribution to the peer review of this work. Primary Handling Editor: Kaitlin McCardle, in collaboration with the Nature Computational Science team. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary sections, Figs. 1–20, Equations 1–4, Algorithm 1 and Table 1.

Peer Review File

Source data

Source Data Fig. 3

Data plotted in Fig. 3a,b, in .csv format.

Source Data Fig. 4

Data plotted in Fig. 4a–d, in .csv format.

Source Data Fig. 5

Data plotted in Fig. 5a,b, in .csv format.

Source Data Fig. 6

Data plotted in Fig. 6a, in .csv format.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kneiding, H., Nova, A. & Balcells, D. Directional multiobjective optimization of metal complexes at the billion-system scale. Nat Comput Sci 4, 263–273 (2024). https://doi.org/10.1038/s43588-024-00616-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s43588-024-00616-5

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing