Scientific Publications

Li S., Chauvin P., Gross O., Bailey M., Jager S.

RECOMB 2026

Introduces a generative language model tailored to mRNA sequence design, enabling the de novo generation of optimised therapeutic mRNA molecules with improved stability and translational efficiency.

Xue Z., Bailey M., Gupta A., Li R., Corrochano-Navarro A., Li S., Kogler-Anele L., Yu Q., Rommelaere H., Van Breedam W., Furtmann N., Batchelor J., Bar-Joseph Z., Jager S.

Journal of Computational Biology, 2025

Applies deep batch active learning to efficiently guide experimental protein structure determination, dramatically reducing the number of experiments needed to build accurate structural models.

Mulyadi A.W., Barker C.G., Kuttae S.B., Wehling L., Rückle T., Singh G., Boucher N., Abdessalem F., Jager S., Siokis A., et al.

EurIPS/NeurIPS SimBioChem 2025

Evaluates whether large time-series foundation models can act as zero-shot surrogates for mechanistic virtual patient models in clinical pharmacology — eliminating costly model re-training. Highlighted and forwarded to Nature.

Mulyadi A.W., Wehling L., Kumar A., Boucher N., Abdessalem F., Jager S., Mosa M.H., et al.

NeurIPS AI4Science 2025

A multi-hop reasoning framework over biomedical knowledge graphs using path-based relational learning, enabling complex question answering across linked biological entities.

Broadbent J., Bailey M., Li M., Paul A., De Lescure L., Chauvin P., Kogler-Anele L., Jangjou Y., Jager S.

NeurIPS Negel 2025

An SE(3)-equivariant graph transformer that predicts aqueous solubility of small molecules with state-of-the-art accuracy by leveraging 3D geometric information invariant to rotation and translation.

Gupta A., Li R., Leal C., Sevvana M., Mackness B.C., Casner R.G., Batchelor J.D., Park A., Bailey M., Kogler Anele L., Jager S., et al.

ICML FM4LS 2025 — Selected Talk

A structure-aware protein language model that integrates 3D structural information into sequence embeddings, achieving superior performance on antibody discovery and optimisation benchmarks.

Xue Z., Bailey M., Gupta A., Li R., Corrochano-Navarro A., Li S., Kogler-Anele L., Yu Q., Bar-Joseph Z., Jager S.

RECOMB 2025 · Lecture Notes in Computer Science, Springer · pp. 17–33

An active learning framework that selectively acquires experimental protein structure data to maximally improve prediction models, minimising the number of expensive crystallography experiments required.

Li X., Amirmoshiri R., Davis C.R., Muthancheri I., de Gombert A., Moayedpour S., Jager S., Rötheli A.R., Jangjou Y.

Industrial & Engineering Chemistry Research, Vol. 64, Issue 13, 2025 · pp. 6825–6837

Combines in silico data generation with probabilistic machine learning to explore reaction mechanisms and build kinetic models for complex chemical processes, accelerating process development.

Li S., Noroozizadeh S., Moayedpour S., Kogler-Anele L., Xue Z., Zheng D., Ulloa Montoya F., Agarwal V., Bar-Joseph Z., Jager S.

Nucleic Acids Research, Vol. 53, Issue 3, 2025

A full-length sequence language model for mRNA that jointly models UTR and coding regions, enabling comprehensive mRNA analysis, stability prediction, and sequence optimisation for therapeutic applications.

Patel R., Moayedpour S., De Lescure L., Kogler-Anele L., Cherney A., Jager S., Jangjou Y.

arXiv, 2025

Distils quantitative chemistry knowledge from LLMs into a prior for Bayesian optimisation, significantly accelerating the search for optimal chemical reaction conditions and reducing laboratory experiments.

Li S., Moayedpour S., Li R., Bailey M., Riahi S., Kogler-Anele L., Miladi M., Miner J., Pertuy F., Zheng D., Wang J., Agarwal V., Bar-Joseph Z., Jager S., et al.

Genome Research, Vol. 34, Issue 7, 2024 · pp. 1027–1035

A large language model pre-trained on codon sequences that improves mRNA vaccine design by predicting translation efficiency and stability, directly supporting the development of next-generation mRNA therapeutics.

Moayedpour S., Corrochano-Navarro A., Sahneh F., Noroozizadeh S., Koetter A., Vymetal J., Kogler-Anele L., Mas P., Jangjou Y., Li S., Bailey M., Matter H., Grebner C., Hessler G., Bar-Joseph Z., Jager S.

ICML AI for Science Workshop 2024

Demonstrates that many-shot in-context learning with LLMs can effectively tackle molecular inverse design, generating molecules with target properties without additional fine-tuning.

Moayedpour S., Broadbent J., Riahi S., Bailey M., Thu H.V., Dobchev D., Balsubramani A., Santos R.N.D., Kogler-Anele L., Corrochano-Navarro A., Li S., Montoya F.U., Agarwal V., Bar-Joseph Z., Jager S.

Bioinformatics, Vol. 40, Issue 7, 2024

Uses large language model embeddings to represent lipid nanoparticles (LNPs) and predict their mRNA transfection efficiency, accelerating the design of superior mRNA delivery vehicles.

Li S., Moayedpour S., Li R., Bailey M., Riahi S., Miladi M., Miner J., Zheng D., Wang J., Balsubramani A., Agarwal V., Bar-Joseph Z., Jager S., et al.

NeurIPS Generative AI and Biology 2023 — Spotlight Paper & Talk

The original CodonBERT spotlight paper: a codon-level BERT model that learns rich representations of mRNA sequences, enabling optimisation of codon usage for improved expression in therapeutic contexts.

Bailey M., Moayedpour S., Li R., Corrochano-Navarro A., Kötter A., Kogler-Anele L., Riahi S., Grebner C., Hessler G., Matter H., Bianciotto M., Mas P., Bar-Joseph Z., Jager S.

eLife, Vol. 12, 2023

Applies deep learning-guided batch active learning to drug discovery, intelligently selecting which compounds to screen experimentally to identify potent candidates while minimising assay costs.

Riahi S., Lee J.H., Sorenson T., Wei S., Jager S., Olfati-Saber R., Zhou Y., Park A., Wendt M., Minoux H., Qiu Y.

Bioinformatics, Vol. 39, Issue 4, 2023

A geometry-aware computational system for comparing protein molecular surfaces using 3D shape descriptors, enabling structure-based drug design and protein function annotation across large protein databases.

Schlichting N., Reinhardt F., Jager S., Schmidt M., Kabisch J.

Oxford Synthetic Biology, Vol. 4, Issue 1, 2019

Systematically optimises the experimental conditions for the ligase cycling reaction (LCR), a key technique in synthetic biology for seamless gene assembly and library construction.

Schmidt M., Hamacher K., Reinhardt F., Lotz T.S., Groher F., Suess B., Jager S.

IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2019

A subgraph isomorphism-based method for comparing RNA secondary structures, identifying structural similarities and functional relationships across large RNA databases with high accuracy.

Groher A., Jager S., Schneider C., Hamacher K., Suess B.

ACS Synthetic Biology, 2019

Applies machine learning to rationally tune riboswitch performance, enabling programmable and quantitatively predictable gene regulation for synthetic biology circuits.

Jager S.

TU Darmstadt, 2018

Doctoral thesis developing computational methods for optimising single molecules and integrating these approaches into automated high-throughput screening workflows for accelerated drug and material discovery.

Jager S., Buss O.

BIOspektrum, Issue 01/18, Springer, 2018

Reviews novel in silico computational methods for establishing green chemistry practices, highlighting how computer-aided approaches can reduce hazardous reagents and waste in chemical synthesis.

Groher F., Bofill-Bosch C., Schneider C., Braun J., Jager S., Geißler K., Hamacher K., Suess B.

Nucleic Acids Research, 2018

Develops and characterises a novel RNA-based genetic switch triggered by the antibiotic ciprofloxacin, expanding the toolkit for ligand-responsive gene regulation in synthetic biology.

Dombrowsky M.J., Jager S., Schiller B., Mayer B.E., Stammler S., Hamacher K.

Journal of Computational Chemistry, 2018

A stream-based R framework for advanced statistical analysis of molecular dynamics simulation trajectories, enabling real-time analysis of large-scale MD simulations without full trajectory storage.

Buss O., Muller D., Jager S., Rudat J., Rabe K.S.

ChemBioChem, 2017

Uses the computational protein design tool FoldX to engineer improved thermostability in an industrially relevant transaminase enzyme, demonstrating in silico-guided biocatalyst engineering.

Jager S., Schiller B., Babel P., Blumenroth M., Strufe T., Hamacher K.

Algorithms for Molecular Biology, 12(1):15, 2017

Introduces Markov model-based algorithms for analysing coarse-grained RNA molecular dynamics, capturing conformational dynamics of RNA structures from connectivity-graph representations.

Gross C., Hamacher K., Schmitz K., Jager S.

ACS Journal of Chemical Information and Modeling, 57(2):243–255, 2017

Computationally investigates how cleavage product accumulation progressively inhibits cutinase activity during enzymatic PET plastic biodegradation, informing enzyme engineering for improved plastic recycling.

Jager S., Schiller B., Strufe T., Hamacher K.

WABI 2016 · Lecture Notes in Computer Science, Vol. 9838, Springer

Conference version presenting the StreAM-Tg algorithm for coarse-grained RNA dynamics analysis, demonstrating its application to benchmark RNA systems and comparison with all-atom MD.

Buss O., Jager S., Dold S.-M., Zimmermann S., Hamacher K., Schmitz K., Rudat J.

PLoS One, 11(1):e0146104

Develops a statistical framework for evaluating high-throughput screening assays targeting enzymatic β-keto ester hydrolysis, enabling reliable hit identification and false positive reduction in enzyme discovery campaigns.

Schiller B., Jager S., Hamacher K., Strufe T.

AlCoB 2015 · Lecture Notes in Computer Science, Vol. 9199, Springer

Introduces a memory-efficient stream-based algorithm for counting network motifs in large dynamic graphs, applicable to biological interaction networks and real-time social network analysis.