Lisa and Kerstin, who now is a junior professor at Universtiy Medicine Göttingen, attended together the BioTechX Europe 2025 conference in Basel. They presented together a poster “Increasing trustworthiness of machine learning-based drug sensitivity prediction with a multivariate random forest approach”, which is also published on ChemRxiv. It was a great opportunity to share our work and interact with fellow researches in academia and industry.
Read more
A new preprint is out! We developed a new fragment-based pipeline to generate novel kinase ligands using subpocket-based docking. In collaboration with the Diederich lab at the Phillips University Marburg, selected compounds were synthesized and tested against the PKA kinase, showing promising activity. We were able to crystallize the most promising hit in complex with PKA.
Thanks for all the co-authors for a great collaboration on this project: Katharina Buchthal, Paula Linh Kramer, Andrea Volkamer and the Diederich lab from the Phillips University Marburg: Dominik Hubach, Johannes Krieger, Gina Bach, Naemi Wagner, Moritz Klopper, Michael Daude and Wibke E. Diederich.
Read more
As part of the 8th Artificial Intelligence in Chemistry Symposium
(AiChem 8 conference website), Raquel, Afnan, and Lisa participated in both the pre-conference workshop and the main conference, contributing to discussions at the intersection of chemistry, data science, and machine learning.
Raquel, Afnan, Lisa, Andrea and Pat organized the pre-conference workshop, which focused on introducing practical machine learning workflows for chemistry. Lisa introduced the basics of Python and Jupyter notebooks, Raquel presented a hands-on notebook on data curation and cleaning, and Afnan explored molecular representations and machine learning models commonly used in cheminformatics. The workshop lectures and notebooks are publicly available
(GitHub repository).
Read more
This year’s EUROPIN Summer School on Drug Design was organized by the Pharmacoinformatics research group in Vienna from 14th-19th of September 2025. Floriane presented a poster on “Integrating structural and morphological fingerprints: Understanding information for pattern identification and better toxicity prediction”. Paula presented her poster on “Active learning for fragment-based kinase inhibitor design”, for which she won the poster prize.
In addition, as part of the EUROPIN PhD program, they both gave their application talks presenting their current research projects. Floriane talked about Uncovering Compound Mode of Action and Patterns via Hierarchical Clustering of Structural and Morphological Profiles. Paula presented her current project on fragment-based active learning for kinase ligand generation. It was a great opportunity to share our work and interact with fellow researches in academia and industry.
Read more
This year’s Gordon’s Seminar and Conference on Computer-Aided Drug Design iteration was held in Portland, Maine. Paula contributed a spotlight presentation and a poster on fragment-based active learning for kinase ligand design.
It was a great experience connecting with researchers from Europe and the U.S., and engaging in insightful discussions with both academic and industry experts.
Read more
Hamza had the opportunity to present part of his PhD research at IOCB Tech in Prague, where a CECAM flagship workshop focused on Quantum Chemistry for Drug Design was hosted. Hamza’s talk covered the topic of adapting machine learning potentials into a physics-based scoring function, which represents a key aspect of his doctoral research. The workshop brought together experts in quantum chemistry and computational drug design, providing an excellent platform for stimulating discussions and valuable feedback. More information can be found on the linkedin post.
Read more
Paula was awarded a Fulbright research scholarship to visit the Chodera Lab at Memorial Sloan Kettering Cancer Center in New York City for four months. She spent the time at the lab to learn about relative binding free energy calculations for kinases and integrating this into her kinase-focused research. She also had the opportunity to get insight into the lab to measure kinase affinities. She enjoyed her time exploring the city, spending time with the lab outside of work and visiting different seminars in New York to connect with US-based researchers.
Read more
This year’s CISPA summer school was co-organized by ELLIS and ELSA from 4th-8th of August 2025 and covered the topic “Trustowrthy AI - Secure and Safe Foundation Models”. Lisa presented her first poster, entitled ‘Trustworthy Machine Learning for Chemical Risk Assessment’ showing the work she is doing for the RADAR project. It was a great opportunity to share our work and interact with fellow researches in academia and industry.
Read more
Big news! Our preprint “Increasing trustworthiness of machine learning-based drug sensitivity prediction with a multivariate random forest approach” is now live on ChemRxiv! We introduce MORGOTH, a trustworthy multivariate random forest model that simultaneously performs classification and regression. Additionally, MORGOTH provides a graph representation of the random forest to address model interpretability, and a cluster analysis of the leaves to measure the dissimilarity of new inputs from the training data to account for its reliability and robustness. While broadly applicable, we focus on anti-cancer drug sensitivity prediction and prioritization. We show that it has state-of-the-art performance while being more trustworhty than any other existing approach.
For more information check out our paper.
Read more
As part of the 13th International Conference on Chemical Structures (ICCS), Tony (Antoine), Floriane, and Afnan presented their recent research at the intersection of cheminformatics, machine learning, and drug discovery. The conference brought together an international community of researchers focused on computational methods for chemical structure representation and analysis.
Tony gave a talk introducing DockM8 and the CADD Vault, highlighting their roles in structure-based drug design and in enabling reproducible and accessible computational workflows for computer-aided drug discovery.
Read more
