{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,7]],"date-time":"2026-04-07T22:00:22Z","timestamp":1775599222593,"version":"3.50.1"},"reference-count":38,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2025,2,25]],"date-time":"2025-02-25T00:00:00Z","timestamp":1740441600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,2,25]],"date-time":"2025-02-25T00:00:00Z","timestamp":1740441600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100023371","name":"Helmholtz Artificial Intelligence Cooperation Unit","doi-asserted-by":"publisher","award":["ZT-I-PF5-23"],"award-info":[{"award-number":["ZT-I-PF5-23"]}],"id":[{"id":"10.13039\/501100023371","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100002347","name":"Bundesministerium f\u00fcr Bildung und Forschung","doi-asserted-by":"publisher","award":["16KISA030"],"award-info":[{"award-number":["16KISA030"]}],"id":[{"id":"10.13039\/501100002347","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Deutsches Zentrum f\u00fcr Neurodegenerative Erkrankungen e.V. (DZNE) in der Helmholtz-Gemeinschaft"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["J Cheminform"],"abstract":"<jats:sec>\n                    <jats:title>Abstract<\/jats:title>\n                    <jats:p>\n                      With the cost\/yield-ratio of drug development becoming increasingly unfavourable, recent work has explored machine learning to accelerate early stages of the development process. Given the current success of deep generative models across domains, we here investigated their application to the property-based proposal of new small molecules for drug development. Specifically, we trained a latent diffusion model\u2014\n                      <jats:italic>DrugDiff<\/jats:italic>\n                      \u2014paired with predictor guidance to generate novel compounds with a variety of desired molecular properties. The architecture was designed to be highly flexible and easily adaptable to future scenarios. Our experiments showed successful generation of unique, diverse and novel small molecules with targeted properties. The code is available at\n                      <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"https:\/\/github.com\/MarieOestreich\/DrugDiff\" ext-link-type=\"uri\">https:\/\/github.com\/MarieOestreich\/DrugDiff<\/jats:ext-link>\n                      .\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Scientific Contribution<\/jats:title>\n                    <jats:p>This work expands the use of generative modelling in the field of drug development from previously introduced models for proteins and RNA to the here presented application to small molecules. With small molecules making up the majority of drugs, but simultaneously being difficult to model due to their elaborate chemical rules, this work tackles a new level of difficulty in comparison to sequence-based molecule generation as is the case for proteins and RNA. Additionally, the demonstrated framework is highly flexible, allowing easy addition or removal of considered molecular properties without the need to retrain the model, making it highly adaptable to diverse research settings and it shows compelling performance for a wide variety of targeted molecular properties.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1186\/s13321-025-00965-x","type":"journal-article","created":{"date-parts":[[2025,2,25]],"date-time":"2025-02-25T04:56:42Z","timestamp":1740459402000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":14,"title":["DrugDiff: small molecule diffusion model with flexible guidance towards molecular properties"],"prefix":"10.1186","volume":"17","author":[{"given":"Marie","family":"Oestreich","sequence":"first","affiliation":[]},{"given":"Erinc","family":"Merdivan","sequence":"additional","affiliation":[]},{"given":"Michael","family":"Lee","sequence":"additional","affiliation":[]},{"given":"Joachim L.","family":"Schultze","sequence":"additional","affiliation":[]},{"given":"Marie","family":"Piraud","sequence":"additional","affiliation":[]},{"given":"Matthias","family":"Becker","sequence":"additional","affiliation":[]}],"member":"297","published-online":{"date-parts":[[2025,2,25]]},"reference":[{"issue":"7958","key":"965_CR1","doi-asserted-by":"publisher","first-page":"673","DOI":"10.1038\/s41586-023-05905-z","volume":"616","author":"AV Sadybekov","year":"2023","unstructured":"Sadybekov AV, Katritch V (2023) Computational approaches streamlining drug discovery. Nature 616(7958):673\u2013685. https:\/\/doi.org\/10.1038\/s41586-023-05905-z","journal-title":"Nature"},{"key":"965_CR2","unstructured":"Deloitte pharma study: Drop-off in returns on R&D investments\u2014sharp decline in peak sales per asset. https:\/\/www2.deloitte.com\/ch\/en\/pages\/press-releases\/articles\/deloitte-pharma-study-drop-off-in-returns-on-r-and-d-investments-sharp-decline-in-peak-sales-per-asset.html (accessed May 08, 2023)."},{"key":"965_CR3","unstructured":"Deloitte. (2023). Seize the digital momentum Measuring the return from pharmaceutical innovation 2022."},{"issue":"3","key":"965_CR4","doi-asserted-by":"publisher","first-page":"191","DOI":"10.1038\/nrd3681","volume":"11","author":"JW Scannell","year":"2012","unstructured":"Scannell JW, Blanckley A, Boldon H, Warrington B (2012) Diagnosing the decline in pharmaceutical R&D efficiency. Nat Rev Drug Discov 11(3):191\u2013200. https:\/\/doi.org\/10.1038\/nrd3681","journal-title":"Nat Rev Drug Discov"},{"issue":"6","key":"965_CR5","doi-asserted-by":"publisher","first-page":"1443","DOI":"10.1039\/d2sc05709c","volume":"14","author":"F Ren","year":"2023","unstructured":"Ren F et al (2023) AlphaFold accelerates artificial intelligence powered drug discovery: efficient discovery of a novel CDK20 small molecule inhibitor. Chem Sci 14(6):1443\u20131452. https:\/\/doi.org\/10.1039\/d2sc05709c","journal-title":"Chem Sci"},{"issue":"9","key":"965_CR6","doi-asserted-by":"publisher","first-page":"1038","DOI":"10.1038\/s41587-019-0224-x","volume":"37","author":"A Zhavoronkov","year":"2019","unstructured":"Zhavoronkov A et al (2019) Deep learning enables rapid identification of potent DDR1 kinase inhibitors. Nat Biotechnol 37(9):1038\u20131040. https:\/\/doi.org\/10.1038\/s41587-019-0224-x","journal-title":"Nat Biotechnol"},{"issue":"7976","key":"965_CR7","doi-asserted-by":"publisher","first-page":"1089","DOI":"10.1038\/s41586-023-06415-8","volume":"620","author":"JL Watson","year":"2023","unstructured":"Watson JL et al (2023) De novo design of protein structure and function with RFdiffusion. Nature 620(7976):1089\u20131100. https:\/\/doi.org\/10.1038\/s41586-023-06415-8","journal-title":"Nature"},{"issue":"8016","key":"965_CR8","doi-asserted-by":"publisher","first-page":"493","DOI":"10.1038\/s41586-024-07487-w","volume":"630","author":"J Abramson","year":"2024","unstructured":"Abramson J et al (2024) Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630(8016):493\u2013500. https:\/\/doi.org\/10.1038\/s41586-024-07487-w","journal-title":"Nature"},{"issue":"2","key":"965_CR9","doi-asserted-by":"publisher","first-page":"392","DOI":"10.1039\/D2DD00115B","volume":"2","author":"J Guo","year":"2023","unstructured":"Guo J et al (2023) Link-INVENT: generative linker design with reinforcement learning. Digital Discovery 2(2):392\u2013408. https:\/\/doi.org\/10.1039\/D2DD00115B","journal-title":"Digital Discovery"},{"issue":"43","key":"965_CR10","doi-asserted-by":"publisher","first-page":"14577","DOI":"10.1039\/d1sc02436a","volume":"12","author":"F Imrie","year":"2021","unstructured":"Imrie F, Hadfield TE, Bradley AR, Deane CM (2021) Deep generative design with 3D pharmacophoric constraints. Chem Sci 12(43):14577\u201314589. https:\/\/doi.org\/10.1039\/d1sc02436a","journal-title":"Chem Sci"},{"key":"965_CR11","doi-asserted-by":"publisher","DOI":"10.1101\/2022.12.01.518682","author":"J Ingraham","year":"2022","unstructured":"Ingraham J et al (2022) Illuminating protein space with a programmable generative model. BioRxiv. https:\/\/doi.org\/10.1101\/2022.12.01.518682","journal-title":"BioRxiv"},{"key":"965_CR12","doi-asserted-by":"publisher","DOI":"10.1002\/wcms.1468","author":"D Schaller","year":"2020","unstructured":"Schaller D et al (2020) Next generation 3D pharmacophore modeling. WIREs Comput Mol Sci. https:\/\/doi.org\/10.1002\/wcms.1468","journal-title":"WIREs Comput Mol Sci"},{"key":"965_CR13","doi-asserted-by":"publisher","DOI":"10.4550\/arxiv.2308.07416","author":"J Torge","year":"2023","unstructured":"Torge J, Harris C, Mathis SV, Lio P (2023) DiffHopp: a graph diffusion model for novel drug design via scaffold hopping. arXiv. https:\/\/doi.org\/10.4550\/arxiv.2308.07416","journal-title":"arXiv"},{"issue":"36","key":"965_CR14","doi-asserted-by":"publisher","first-page":"28","DOI":"10.1021\/cen-10136-cover","volume":"101","author":"L Howes","year":"2023","unstructured":"Howes L (2023) Is this a golden age of small-molecule drug discovery? C&EN Global Enterp 101(36):28\u201332. https:\/\/doi.org\/10.1021\/cen-10136-cover","journal-title":"C&EN Global Enterp"},{"key":"965_CR15","doi-asserted-by":"publisher","unstructured":"R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer. (2021). High-resolution image synthesis with latent diffusion models. arXiv. https:\/\/doi.org\/10.48550\/arxiv.2112.10752.","DOI":"10.48550\/arxiv.2112.10752"},{"key":"965_CR16","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-024-00843-5","author":"Y Du","year":"2024","unstructured":"Du Y et al (2024) Machine learning-aided generative molecular design. Nat Mach Intell. https:\/\/doi.org\/10.1038\/s42256-024-00843-5","journal-title":"Nat Mach Intell"},{"issue":"19","key":"965_CR17","doi-asserted-by":"publisher","first-page":"7238","DOI":"10.1021\/acs.jcim.4c01107","volume":"64","author":"A Alakhdar","year":"2024","unstructured":"Alakhdar A, Poczos B, Washburn N (2024) Diffusion models in de novo drug design. J Chem Inf Model 64(19):7238\u20137256. https:\/\/doi.org\/10.1021\/acs.jcim.4c01107","journal-title":"J Chem Inf Model"},{"issue":"1","key":"965_CR18","doi-asserted-by":"publisher","first-page":"2657","DOI":"10.1038\/s41467-024-46569-1","volume":"15","author":"L Huang","year":"2024","unstructured":"Huang L et al (2024) A dual diffusion model enables 3D molecule generation and lead optimization based on target pockets. Nat Commun 15(1):2657. https:\/\/doi.org\/10.1038\/s41467-024-46569-1","journal-title":"Nat Commun"},{"key":"965_CR19","doi-asserted-by":"publisher","first-page":"110992","DOI":"10.1016\/j.isci.2024.110992","volume":"27","author":"Y Luo","year":"2024","unstructured":"Luo Y et al (2024) Text-guided small molecule generation via diffusion model. iScience 27:110992. https:\/\/doi.org\/10.1016\/j.isci.2024.110992","journal-title":"iScience"},{"key":"965_CR20","doi-asserted-by":"publisher","DOI":"10.48550\/arxiv.2305.13266","author":"B Qiang","year":"2023","unstructured":"Qiang B et al (2023) Coarse-to-Fine: a hierarchical diffusion model for molecule generation in 3D. Int Conf Machine Learn. https:\/\/doi.org\/10.48550\/arxiv.2305.13266","journal-title":"Int Conf Machine Learn"},{"issue":"19","key":"965_CR21","doi-asserted-by":"publisher","first-page":"5996","DOI":"10.1021\/acs.jcim.3c00667","volume":"63","author":"NT Runcie","year":"2023","unstructured":"Runcie NT, Mey ASJS (2023) SILVR: guided diffusion for molecule generation. J Chem Inf Model 63(19):5996\u20136005. https:\/\/doi.org\/10.1021\/acs.jcim.3c00667","journal-title":"J Chem Inf Model"},{"key":"965_CR22","doi-asserted-by":"publisher","DOI":"10.48550\/arxiv.2305.01140","author":"M Xu","year":"2023","unstructured":"Xu M, Powers A, Dror R, Ermon S, Leskovec J (2023) Geometric latent diffusion models for 3D molecule generation. Int Conf Machine Learn. https:\/\/doi.org\/10.48550\/arxiv.2305.01140","journal-title":"Int Conf Machine Learn"},{"key":"965_CR23","first-page":"5777","volume":"162","author":"P Eckmann","year":"2022","unstructured":"Eckmann P, Sun K, Zhao B, Feng M, Gilson MK, Yu R (2022) LIMO: Latent inceptionism for targeted molecule generation presented at the ICML. Proc Machine Learn Res 162:5777","journal-title":"Proc Machine Learn Res"},{"issue":"11","key":"965_CR24","doi-asserted-by":"publisher","first-page":"2324","DOI":"10.1021\/acs.jcim.5b00559","volume":"55","author":"T Sterling","year":"2015","unstructured":"Sterling T, Irwin JJ (2015) ZINC 15\u2013ligand discovery for everyone. J Chem Inf Model 55(11):2324\u20132337. https:\/\/doi.org\/10.1021\/acs.jcim.5b00559","journal-title":"J Chem Inf Model"},{"issue":"12","key":"965_CR25","doi-asserted-by":"publisher","first-page":"6065","DOI":"10.1021\/acs.jcim.0c00675","volume":"60","author":"JJ Irwin","year":"2020","unstructured":"Irwin JJ et al (2020) ZINC20-A free ultralarge-scale chemical database for ligand discovery. J Chem Inf Model 60(12):6065\u20136073. https:\/\/doi.org\/10.1021\/acs.jcim.0c00675","journal-title":"J Chem Inf Model"},{"issue":"3","key":"965_CR26","doi-asserted-by":"publisher","first-page":"1096","DOI":"10.1021\/acs.jcim.8b00839","volume":"59","author":"N Brown","year":"2019","unstructured":"Brown N, Fiscato M, Segler MHS, Vaucher AC (2019) Guacamol: benchmarking models for de novo molecular design. J Chem Inf Model 59(3):1096\u20131108. https:\/\/doi.org\/10.1021\/acs.jcim.8b00839","journal-title":"J Chem Inf Model"},{"issue":"20","key":"965_CR27","doi-asserted-by":"publisher","first-page":"3714","DOI":"10.1021\/jm000942e","volume":"43","author":"P Ertl","year":"2000","unstructured":"Ertl P, Rohde B, Selzer P (2000) Fast calculation of molecular polar surface area as a sum of fragment-based contributions and its application to the prediction of drug transport properties. J Med Chem 43(20):3714\u20133717. https:\/\/doi.org\/10.1021\/jm000942e","journal-title":"J Med Chem"},{"key":"965_CR28","doi-asserted-by":"publisher","DOI":"10.1186\/1758-2946-1-8","author":"P Ertl","year":"2009","unstructured":"Ertl P, Schuffenhauer A (2009) Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions. J Cheminform. https:\/\/doi.org\/10.1186\/1758-2946-1-8","journal-title":"J. Cheminform"},{"issue":"12","key":"965_CR29","doi-asserted-by":"publisher","first-page":"2615","DOI":"10.1021\/jm020017n","volume":"45","author":"DF Veber","year":"2002","unstructured":"Veber DF, Johnson SR, Cheng H-Y, Smith BR, Ward KW, Kopple KD (2002) Molecular properties that influence the oral bioavailability of drug candidates. J Med Chem 45(12):2615\u20132623. https:\/\/doi.org\/10.1021\/jm020017n","journal-title":"J Med Chem"},{"issue":"2","key":"965_CR30","doi-asserted-by":"publisher","first-page":"90","DOI":"10.1038\/nchem.1243","volume":"4","author":"GR Bickerton","year":"2012","unstructured":"Bickerton GR, Paolini GV, Besnard J, Muresan S, Hopkins AL (2012) Quantifying the chemical beauty of drugs. Nat Chem 4(2):90\u201398. https:\/\/doi.org\/10.1038\/nchem.1243","journal-title":"Nat Chem"},{"issue":"5","key":"965_CR31","doi-asserted-by":"publisher","first-page":"868","DOI":"10.1021\/ci990307l","volume":"39","author":"SA Wildman","year":"1999","unstructured":"Wildman SA, Crippen GM (1999) Prediction of physicochemical parameters by atomic contributions. J Chem Inf Comput Sci 39(5):868\u2013873. https:\/\/doi.org\/10.1021\/ci990307l","journal-title":"J Chem Inf Comput Sci"},{"key":"965_CR32","unstructured":"M. J. Kusner, B. Paige, and J. M. Hern\u00e1ndez-Lobato. (2017). Grammar Variational Autoencoder.pdf,\u201d Proceedings of the 34th International Conference on Machine Learning."},{"issue":"1","key":"965_CR33","doi-asserted-by":"publisher","first-page":"8799","DOI":"10.1038\/s41598-023-35648-w","volume":"13","author":"E Mazuz","year":"2023","unstructured":"Mazuz E, Shtar G, Shapira B, Rokach L (2023) Molecule generation using transformers and policy gradient reinforcement learning. Sci Rep 13(1):8799. https:\/\/doi.org\/10.1038\/s41598-023-35648-w","journal-title":"Sci Rep"},{"issue":"2","key":"965_CR34","doi-asserted-by":"publisher","first-page":"1060","DOI":"10.1021\/acs.jctc.0c01006","volume":"17","author":"D Santos-Martins","year":"2021","unstructured":"Santos-Martins D, Solis-Vasquez L, Tillack AF, Sanner MF, Koch A, Forli S (2021) Accelerating AutoDock4 with GPUs and gradient-based local search. J Chem Theory Comput 17(2):1060\u20131073. https:\/\/doi.org\/10.1021\/acs.jctc.0c01006","journal-title":"J Chem Theory Comput"},{"key":"965_CR35","unstructured":"D. P. Kingma and M. (2013). Welling. Auto-encoding variational bayes. arXiv"},{"issue":"7","key":"965_CR36","doi-asserted-by":"publisher","first-page":"1757","DOI":"10.1021\/ci3001277","volume":"52","author":"JJ Irwin","year":"2012","unstructured":"Irwin JJ, Sterling T, Mysinger MM, Bolstad ES, Coleman RG (2012) ZINC: a free tool to discover chemistry for biology. J Chem Inf Model 52(7):1757\u20131768. https:\/\/doi.org\/10.1021\/ci3001277","journal-title":"J Chem Inf Model"},{"issue":"1\u20133","key":"965_CR37","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1016\/S0169-409X(96)00423-1","volume":"46","author":"CA Lipinski","year":"2001","unstructured":"Lipinski CA, Lombardo F, Dominy BW, Feeney PJ (2001) Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 46(1\u20133):3\u201326. https:\/\/doi.org\/10.1016\/S0169-409X(96)00423-1","journal-title":"Adv Drug Deliv Rev"},{"key":"965_CR38","first-page":"6480","volume":"33","author":"J Ho","year":"2020","unstructured":"Ho J (2020) Denoising diffusion probabilistic models. Adv Neural Inform Process Syst 33:6480","journal-title":"Adv Neural Inform Process Syst"}],"container-title":["Journal of Cheminformatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s13321-025-00965-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s13321-025-00965-x\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s13321-025-00965-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,2,25]],"date-time":"2025-02-25T04:56:53Z","timestamp":1740459413000},"score":1,"resource":{"primary":{"URL":"https:\/\/jcheminf.biomedcentral.com\/articles\/10.1186\/s13321-025-00965-x"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,2,25]]},"references-count":38,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["965"],"URL":"https:\/\/doi.org\/10.1186\/s13321-025-00965-x","relation":{"has-preprint":[{"id-type":"doi","id":"10.1101\/2024.07.17.603873","asserted-by":"object"}]},"ISSN":["1758-2946"],"issn-type":[{"value":"1758-2946","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,2,25]]},"assertion":[{"value":"24 October 2024","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"27 January 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"25 February 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"An issue that needs addressing is the dual-use problem of models like the one presented here. Generating molecules that manifest desired properties has many beneficial use cases, for example in the medical context or materials sciences. However, it also bears the inherent risk of being misused for malicious intent, for example to generate molecules that maximise toxicity. The model described in this work is solely intended to be used in a beneficial context and in accordance with the law.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"The authors declare no competing interests.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"23"}}