Lucas Arnoldt

Lucas Arnoldt

Auto-generated profile, based on publicly available information

Lucas Arnoldt

Digital Health Center, Berlin Institute of Health (BIH) and Charité University Medicine, Berlin, Germany

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 September 2024 in Nature Biotechnology
Verified
Journal Article
Open access
PDF file
SLSidney Lyayuga Lisanza
Sidney Lyayuga Lisanza
JGJacob Merle Gershon
Jacob Merle Gershon
STSamuel W. K. Tipps
Samuel W. K. Tipps
JSJeremiah Nelson Sims
Jeremiah Nelson Sims
LALucas Arnoldt
Lucas Arnoldt
DBDavid Baker
David Baker
Protein denoising diffusion probabilistic models are used for the de novo generation of protein backbones but are limited in their ability to guide generation of proteins with sequence-specific attributes and functional properties. To overcome this limitation, we developed ProteinGenerator (PG), a sequence space diffusion model based on RoseTTAFold that simultaneously generates protein sequences and structures. Beginning from a noised sequence representation, PG generates sequence and structure pairs by iterative denoising, guided by desired sequence and structural protein attributes. We designed thermostable proteins with varying amino acid compositions and internal sequence repeats and cage bioactive peptides, such as melittin. By averaging sequence logits between diffusion trajectories with distinct structural constraints, we designed multistate parent–child protein triples in which the same sequence folds to different supersecondary structures when intact in the parent versus split into two child domains. PG design trajectories can be guided by experimental sequence–activity data, providing a general approach for integrated computational and experimental optimization of protein function.

10 May 2023 in bioRxiv
Verified
Preprint
Open access
PDF file
SLSidney Lyayuga Lisanza
Sidney Lyayuga Lisanza
JGJake Merle Gershon
Jake Merle Gershon
STSam Tipps
Sam Tipps
LALucas Arnoldt
Lucas Arnoldt
SHSamuel Hendel
Samuel Hendel
DBDavid Baker
David Baker
Protein denoising diffusion probabilistic models (DDPMs) show great promise in the de novo generation of protein backbones but are limited in their inability to guide generation of proteins with sequence specific attributes and functional properties. To overcome this limitation, we develop ProteinGenerator, a sequence space diffusion model based on RoseTTAfold that simultaneously generates protein sequences and structures. Beginning from random amino acid sequences, our model generates sequence and structure pairs by iterative denoising, guided by any desired sequence and structural protein attributes. To explore the versatility of this approach, we designed proteins enriched for specific amino acids, with internal sequence repeats, with masked bioactive peptides, with state dependent structures, and with key sequence features of specific protein families. ProteinGenerator readily generates sequence-structure pairs satisfying the input conditioning (sequence and/or structural) criteria, and experimental validation showed that the designs were monomeric by size exclusion chromatography (SEC), had the desired secondary structure content by circular dichroism (CD), and were thermostable up to 95°C. By enabling the simultaneous optimization of both sequence and structure, ProteinGenerator allows for the design of functional proteins with specific sequence and structural attributes, and paves the way for protein function optimization by active learning on sequence-activity datasets.

7 October 2024 in Nature Biotechnology
Verified
Journal Article
Open access
PDF file
SLSidney Lyayuga Lisanza
Sidney Lyayuga Lisanza
JGJacob Merle Gershon
Jacob Merle Gershon
STSamuel W. K. Tipps
Samuel W. K. Tipps
JSJeremiah Nelson Sims
Jeremiah Nelson Sims
LALucas Arnoldt
Lucas Arnoldt
DBDavid Baker
David Baker

12 June 2025 in bioRxiv
Verified
Preprint
Open access
LALucas Arnoldt
Lucas Arnoldt
JBJulius Upmeier zu Belzen
Julius Upmeier zu Belzen
LHLuis Herrmann
Luis Herrmann
KNKhue Nguyen
Khue Nguyen
FTFabian Theis
Fabian Theis
RERoland Eils
Roland Eils
Multi-omics technologies allow for a detailed characterization of cell types and states across multiple omics layers, helping to identify features that differentiate biological conditions, such as chemical or CRISPR-based perturbations. However, current tools employing variational inference on single-cell datasets, including methods for paired and mosaic integration, transfer learning, and modality imputation, typically act as black boxes. This lack of interpretability makes it challenging to evaluate whether biological variation is preserved, which can compromise downstream analyses. Here, we introduce NetworkVI, a sparse deep generative model designed for the integration and interpretation of multimodal single-cell data. NetworkVI utilizes biological prior knowledge as an inductive bias, specifically it relies on gene-gene interactions inferred from topologically associated domains and structured ontologies like the Gene Ontology to aggregate gene embeddings to cell embeddings, enhancing the interpretability at the gene and subcellular level. While achieving state-of-the-art data integration, modality imputation, and cell label transfer via query-to-reference mapping benchmarks across bimodal and trimodal datasets, NetworkVI additionally excels in providing biologically meaningful modality- and cell type-specific interpretations. NetworkVI aids researchers in identifying associations between genes and biological processes and uncovers immune evasion mechanisms in a Perturb CITE-seq dataset of melanoma cells. NetworkVI will support researchers in interpreting cellular disease mechanisms, guiding biomarker discovery, and ultimately aiding the development of targeted therapies in large-scale single-cell multimodal atlases. NetworkVI is available at http://github.com/LArnoldt/networkVI .

19 March 2026 in The Lancet Digital Health
Verified
Journal Article
Open access
TBThore Buergel
Thore Buergel
LLLukas Loock
Lukas Loock
JSJakob Steinfeldt
Jakob Steinfeldt
NKNoel Kronenberg
Noel Kronenberg
LHLaura Hoffmann
Laura Hoffmann
RERoland Eils
Roland Eils
Early detection of individuals at high risk of disease onset is crucial for health-care systems to cope with changing demographics and an ever-increasing patient population. Images of the retinal fundus are a non-invasive, low-cost examination routinely collected and potentially scalable beyond ophthalmology. Previous work showed the potential of retinal images for risk assessment for some diseases, but it remains unclear whether this potential extends to a broader range of human diseases. We aimed to systematically assess the extent to which retinal fundus photographs can predict incident disease onset across the human phenome, and to benchmark their added value beyond readily available patient characteristics. In this modelling study using data from population-based cohorts, we extended a retinal foundation model (RETFound) to systematically explore the predictive potential of retinal images as a screening strategy for disease onset across 752 incident diseases in 61 256 individuals (33 285 [54%] females and 27 971 [46%] males; median age 58 years [IQR 50-63]) from the UK Biobank cohort. Participants had retinal images collected at baseline (Dec 7, 2009, to July 21, 2010) and were linked to routinely collected hospital and death records in the UK. External validation was performed in 7248 individuals (median age 67 years [IQR 62-73]) from the EPIC-Norfolk Eye Study. Predictive improvements were investigated by extracting image attributions from risk models and performing genome-wide association studies. We showed improved discriminative performance compared with readily available patient characteristics for 306 (41%) of the 752 investigated disease endpoints, including 280 outside of ophthalmology. Retinal information did not improve the prediction for the onset of cardiovascular diseases compared with established primary prevention scores. Predictive improvements were attributable to retinal vascularisation patterns and less obvious features, such as eye colour or lens morphology. Genetic findings highlighted commonalities between eye-derived risk estimates and complex disorders: across 84 retinal risk phenotypes, we identified 1385 genome-wide statistically significant variant associations across 178 loci, including a low-frequency missense variant in IMPA1 (rs204781; minor allele frequency 2·0%) associated with decreased risk estimates across 48 diseases, with the strongest association observed for iron deficiency anaemia (β=-0·16; p<7·2 × 10-16). We present one of the first comprehensive evaluations of predictive information derived from retinal fundus photographs, illustrating the potential and limitations of readily accessible and low-cost retinal images for risk assessment across common and rare diseases. Our findings show the potential of retinal images to complement screening strategies more widely, but also demonstrate the need for rigorous benchmarking and disease-agnostic efforts to design cost-efficient screening strategies to improve population health. Charité - Universitätsmedizin Berlin.

30 July 2026 in medRxiv
Verified
Preprint
Open access
PDF file
JBJulius Upmeier zu Belzen
Julius Upmeier zu Belzen
LALucas Arnoldt
Lucas Arnoldt
NHNoah Hollmann
Noah Hollmann
LHLuis Hermann
Luis Hermann
KNKhue M. Nguyen
Khue M. Nguyen
RERoland Eils
Roland Eils
Genetic prediction of complex phenotypes typically relies on additive linear models, which scale well but cannot capture non-additive effects or deeply integrate molecular and clinical data. Domain-specific neural networks have driven advances in images, text, and other modalities, but genome-scale neural networks remain challenging because genotypes are sparse and high-dimensional, effective sample sizes are limited, and generic architectures lack interpretability. Here, we introduce the omnigenic neural network, a biologically structured architecture inspired by the omnigenic model of complex traits. The model learns hierarchical representations of biological processes, accommodates multimodal inputs, supports transfer learning, and enables multitask prediction. Models trained in the UK Biobank and evaluated in the All of Us cohort for ischemic heart disease, type 2 diabetes, and schizophrenia outperformed published PGS Catalog and PRS-CSx scores. A multitask model trained across 36 cardiovascular endpoints further outperformed corresponding single-phenotype models and baselines. The architecture provides systems-level interpretability by quantifying the contributions of biological processes, which were consistent with established disease mechanisms. It also captures non-linear interactions between variants. Analysis of these interactions using Integrated Hessians revealed patterns concordant with previously reported epistatic associations. Together, these findings establish the omnigenic neural network as a flexible framework for interpretable, multimodal, and multitask genomic prediction.