Roman Thielemann
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Roman Thielemann has a total of 117 co-authors. Here are the top co-authors by number of shared publications:
- Peer Bork
Molecular Systems Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany
6 shared publications - Oleksandr M. Maistrenko
Department of Marine Microbiology and Biogeochemistry, Royal Netherlands Institute for Sea Research, Texel, The Netherlands
5 shared publications - Thea Van Rossum
European Molecular Biology Laboratory, Molecular Systems Biology Unit , 69117 Heidelberg
5 shared publications - Thomas Sebastian Benedikt Schmidt
APC Microbiome and School of Medicine, University College Cork, Cork, Ireland
5 shared publications - Torben Hansen
Novo Nordisk Foundation Center for Basic Metabolic Research, Faculty of Health and Medical Sciences, University of Copenhagen, Blegdamsvej 3B, 2200 Copenhagen N, Denmark
5 shared publications - Wasiu Akanni
European Molecular Biology Laboratory, Structural and Computational Biology Unit
5 shared publications
Latest Publications
- Open Access
- 14 January 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Identification of modifiable plasma protein markers of cardiometabolic risk in children and adolescents with obesity
Pediatric obesity is linked to multi-organ inflammation and an increased risk of cardiometabolic and steatotic liver disease. To identify circulating biomarkers of cardiometabolic risk, we performed proximity extension assay proteomics, to quantify 149 inflammation- and cardiovascular-related proteins in a cross-sectional study of 4024 children and adolescents (2377 with obesity and 1647 with normal weight). We identified protein signatures linked to obesity, dyslipidemia, insulin resistance, hyperglycemia, hypertension, and related cardiometabolic phenotypes. Using machine learning, a three-protein panel (CDCP1, FGF21, HAOX1) combined with liver enzymes improved prediction of steatotic liver disease versus liver enzymes alone (receiver operating characteristic–area under the curve (ROC–AUC) = 0.83 vs. 0.77; DeLong’s test, P < 0.05). During a 1-year non-pharmacological obesity intervention (n = 184), reductions in adiposity were associated with decreased inflammatory cytokines (including CDCP1, FGF21), which correlated with improvements in cardiometabolic risk profiles. Here we show that circulating proteomic signatures may mediate obesity-related cardiometabolic risk in youth.
- 14 January 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Open AccessIdentification of modifiable plasma protein markers of cardiometabolic risk in children and adolescents with obesity
Pediatric obesity is linked to multi-organ inflammation and an increased risk of cardiometabolic and steatotic liver disease. To identify circulating biomarkers of cardiometabolic risk, we performed proximity extension assay proteomics, to quantify 149 inflammation- and cardiovascular-related proteins in a cross-sectional study of 4024 children and adolescents (2377 with obesity and 1647 with normal weight). We identified protein signatures linked to obesity, dyslipidemia, insulin resistance, hyperglycemia, hypertension, and related cardiometabolic phenotypes. Using machine learning, a three-protein panel (CDCP1, FGF21, HAOX1) combined with liver enzymes improved prediction of steatotic liver disease versus liver enzymes alone (receiver operating characteristic–area under the curve (ROC–AUC) = 0.83 vs. 0.77; DeLong’s test, P < 0.05). During a 1-year non-pharmacological obesity intervention (n = 184), reductions in adiposity were associated with decreased inflammatory cytokines (including CDCP1, FGF21), which correlated with improvements in cardiometabolic risk profiles. Here we show that circulating proteomic signatures may mediate obesity-related cardiometabolic risk in youth.
- Open Access
- 14 December 2025
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Complex genetic effects linked to plasma protein abundance in the UK Biobank
Understanding genetic associations of proteins is important for studying the molecular effect of genetic variation. A key component of this is to understand the role of complex genetic effects such as dominance and epistasis that are associated with plasma proteins. Therefore, we develop EIR-auto-GP, a deep learning-based approach, to identify complex effects that are associated with protein quantitative trait loci (pQTLs). Applying this method to the UK Biobank proteomics cohort of 48,594 individuals, we identify 123 proteins that are correlated with non-linear covariates and 15 with genetic dominance and epistasis. We uncover a novel interaction between the ABO and FUT3 loci and demonstrate dominance effects of the ABO locus on plasma levels of pathogen recognition receptors CD209 and CLEC4M. Furthermore, we replicate these findings and the methodology across Olink and mass spectrometry-based cohorts. Our approach presents a systematic, large-scale attempt to identify complex effects of plasma protein levels.
- 14 December 2025
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Open AccessComplex genetic effects linked to plasma protein abundance in the UK Biobank
Understanding genetic associations of proteins is important for studying the molecular effect of genetic variation. A key component of this is to understand the role of complex genetic effects such as dominance and epistasis that are associated with plasma proteins. Therefore, we develop EIR-auto-GP, a deep learning-based approach, to identify complex effects that are associated with protein quantitative trait loci (pQTLs). Applying this method to the UK Biobank proteomics cohort of 48,594 individuals, we identify 123 proteins that are correlated with non-linear covariates and 15 with genetic dominance and epistasis. We uncover a novel interaction between the ABO and FUT3 loci and demonstrate dominance effects of the ABO locus on plasma levels of pathogen recognition receptors CD209 and CLEC4M. Furthermore, we replicate these findings and the methodology across Olink and mass spectrometry-based cohorts. Our approach presents a systematic, large-scale attempt to identify complex effects of plasma protein levels.
- Open Access
- 31 October 2025
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 54 (D1) , D826-D834
Metalog: curated and harmonised contextual data for global metagenomics samples
Metagenomic sequencing enables the in-depth study of microbes and their functions in humans, animals, and the environment. While sequencing data is deposited in public databases, the associated contextual data is often not complete and needs to be retrieved from primary publications. This lack of access to sample-level metadata like clinical data or in situ observations impedes cross-study comparisons and meta-analyses. We therefore created the Metalog database, a repository of manually curated metadata for metagenomics samples across the globe. It contains 80 423 samples from humans (including 66 527 of the gut microbiome), 10 744 animal samples, 5547 ocean water samples, and 23 455 samples from other environmental habitats such as soil, sediment, or fresh water. Samples have been consistently annotated for a set of habitat-specific core features, such as demographics, disease status, and medication for humans; host species and captivity status for animals; and filter sizes and salinity for marine samples. Additionally, all original metadata is provided in tabular form, simplifying focused studies e.g. into nutrient concentrations. Pre-computed taxonomic profiles facilitate rapid data exploration, while links to the SPIRE database enable genome-based analyses. The database is freely available for browsing and download at https://metalog.embl.de/.
- 31 October 2025
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 54 (D1) , D826-D834
Open AccessMetalog: curated and harmonised contextual data for global metagenomics samples
Metagenomic sequencing enables the in-depth study of microbes and their functions in humans, animals, and the environment. While sequencing data is deposited in public databases, the associated contextual data is often not complete and needs to be retrieved from primary publications. This lack of access to sample-level metadata like clinical data or in situ observations impedes cross-study comparisons and meta-analyses. We therefore created the Metalog database, a repository of manually curated metadata for metagenomics samples across the globe. It contains 80 423 samples from humans (including 66 527 of the gut microbiome), 10 744 animal samples, 5547 ocean water samples, and 23 455 samples from other environmental habitats such as soil, sediment, or fresh water. Samples have been consistently annotated for a set of habitat-specific core features, such as demographics, disease status, and medication for humans; host species and captivity status for animals; and filter sizes and salinity for marine samples. Additionally, all original metadata is provided in tabular form, simplifying focused studies e.g. into nutrient concentrations. Pre-computed taxonomic profiles facilitate rapid data exploration, while links to the SPIRE database enable genome-based analyses. The database is freely available for browsing and download at https://metalog.embl.de/.
- Open Access
Metalog: curated and harmonised contextual data for global metagenomics samples
Metagenomic sequencing enables the in-depth study of microbes and their functions in humans, animals and the environment. While sequencing data is deposited in public databases, the associated contextual data is often not complete and needs to be retrieved from primary publications. This lack of access to sample-level metadata like clinical data or in situ observations impedes cross-study comparisons and meta-analyses. We therefore created the Metalog database, a repository of manually curated metadata for metagenomics samples across the globe. It contains 73,082 samples from humans (including 58,506 of the gut microbiome), 10,703 animal samples, 5,146 ocean water samples, and 21,802 samples from other environmental habitats such as soil, sediment, or fresh water. Samples have been consistently annotated for a set of habitat-specific core features, such as demographics, disease status and medication for humans, host species and captivity status for animals, and filter sizes and salinity for marine samples. Additionally, all original metadata is provided in tabular form, simplifying focused studies e.g. into nutrient concentrations. Pre-computed taxonomic profiles facilitate rapid data exploration, while links to the SPIRE database enable genome-based analyses. The database is freely available for browsing and download at https://metalog.embl.de/ .
Open AccessMetalog: curated and harmonised contextual data for global metagenomics samples
Metagenomic sequencing enables the in-depth study of microbes and their functions in humans, animals and the environment. While sequencing data is deposited in public databases, the associated contextual data is often not complete and needs to be retrieved from primary publications. This lack of access to sample-level metadata like clinical data or in situ observations impedes cross-study comparisons and meta-analyses. We therefore created the Metalog database, a repository of manually curated metadata for metagenomics samples across the globe. It contains 73,082 samples from humans (including 58,506 of the gut microbiome), 10,703 animal samples, 5,146 ocean water samples, and 21,802 samples from other environmental habitats such as soil, sediment, or fresh water. Samples have been consistently annotated for a set of habitat-specific core features, such as demographics, disease status and medication for humans, host species and captivity status for animals, and filter sizes and salinity for marine samples. Additionally, all original metadata is provided in tabular form, simplifying focused studies e.g. into nutrient concentrations. Pre-computed taxonomic profiles facilitate rapid data exploration, while links to the SPIRE database enable genome-based analyses. The database is freely available for browsing and download at https://metalog.embl.de/ .
- Open Access
Obesity- and age-dependent genetic regulation of the plasma proteome in children and adolescents
The genetic regulation of the plasma proteome has been extensively studied in adult populations, yet protein quantitative trait loci (pQTL) studies in children and adolescents remain largely unexplored. Here, we mapped pQTLs for 178 plasma proteins measured using affinity-based proteomics in 3,853 Danish children and adolescents (44.1% boys; median age of 11.6 years) from the HOLBAEK Study. We aimed to identify context-dependent pQTLs, where genetic variant-protein associations varied across biological contexts such as obesity, puberty, and sex. We further investigated pQTL variation across the lifespan by comparing our findings to pQTL data from UK Biobank. In the HOLBAEK Study, we identified 1,328 independent, genome-wide significant associations for 178 proteins, replicating previously identified pQTLs in adult cohorts. We identified obesity-dependent cis -pQTLs for IL-1ra, TRANCE, and PIgR and found nominally significant evidence for puberty- and sex-dependent genetic differences. Comparative analysis revealed age-dependent differences in cis -pQTLs for 30 proteins between children and adolescents of the HOLBAEK Study (aged 4 to 20 years) and adults from the UK Biobank (aged 40 to 70 years). Through one-sample Mendelian randomization analysis, we identified 70 protein–cardiometabolic trait links at nominal significance, highlighting the potential causal role of proteins in cardiometabolic health in children and adolescents.
Open AccessObesity- and age-dependent genetic regulation of the plasma proteome in children and adolescents
The genetic regulation of the plasma proteome has been extensively studied in adult populations, yet protein quantitative trait loci (pQTL) studies in children and adolescents remain largely unexplored. Here, we mapped pQTLs for 178 plasma proteins measured using affinity-based proteomics in 3,853 Danish children and adolescents (44.1% boys; median age of 11.6 years) from the HOLBAEK Study. We aimed to identify context-dependent pQTLs, where genetic variant-protein associations varied across biological contexts such as obesity, puberty, and sex. We further investigated pQTL variation across the lifespan by comparing our findings to pQTL data from UK Biobank. In the HOLBAEK Study, we identified 1,328 independent, genome-wide significant associations for 178 proteins, replicating previously identified pQTLs in adult cohorts. We identified obesity-dependent cis -pQTLs for IL-1ra, TRANCE, and PIgR and found nominally significant evidence for puberty- and sex-dependent genetic differences. Comparative analysis revealed age-dependent differences in cis -pQTLs for 30 proteins between children and adolescents of the HOLBAEK Study (aged 4 to 20 years) and adults from the UK Biobank (aged 40 to 70 years). Through one-sample Mendelian randomization analysis, we identified 70 protein–cardiometabolic trait links at nominal significance, highlighting the potential causal role of proteins in cardiometabolic health in children and adolescents.
- Open Access
Identification of modifiable plasma protein markers of cardiometabolic risk in children and adolescents with obesity
Pediatric obesity is associated with multi-organ inflammation and an increased risk of cardiometabolic disease. To identify novel biomarkers of early-onset obesity-related cardiometabolic risk, we performed targeted proteomics, analyzing 149 proteins related to inflammation and cardiovascular disease in a cross-sectional sample of 2,377 children and adolescents with overweight/obesity and 1,647 with normal weight. In addition, we analyzed 64 circulating inflammation-related proteins in 184 children and adolescents, who participated in a 1-year family-based obesity management program. We identified a three plasma protein panel (CDCP1, FGF21 and HAOX1) that improved prediction of hepatic steatosis, compared with the predictive value of liver enzymes alone. In the prospective trial, we found that the non-pharmacological intervention induced a decline in circulating inflammatory cytokine levels, some of which were linked to improvements in cardiometabolic phenotypes. Our results may suggest that circulating proteomic signatures may mediate the associations between pediatric obesity and cardiometabolic risk.
Open AccessIdentification of modifiable plasma protein markers of cardiometabolic risk in children and adolescents with obesity
Pediatric obesity is associated with multi-organ inflammation and an increased risk of cardiometabolic disease. To identify novel biomarkers of early-onset obesity-related cardiometabolic risk, we performed targeted proteomics, analyzing 149 proteins related to inflammation and cardiovascular disease in a cross-sectional sample of 2,377 children and adolescents with overweight/obesity and 1,647 with normal weight. In addition, we analyzed 64 circulating inflammation-related proteins in 184 children and adolescents, who participated in a 1-year family-based obesity management program. We identified a three plasma protein panel (CDCP1, FGF21 and HAOX1) that improved prediction of hepatic steatosis, compared with the predictive value of liver enzymes alone. In the prospective trial, we found that the non-pharmacological intervention induced a decline in circulating inflammatory cytokine levels, some of which were linked to improvements in cardiometabolic phenotypes. Our results may suggest that circulating proteomic signatures may mediate the associations between pediatric obesity and cardiometabolic risk.
- Open Access
- 2 January 2025
- Published by Elsevier in EBioMedicine
- Vol. 112, 105537
Genetic factors shaping the plasma lipidome and the relations to cardiometabolic risk in children and adolescents
- 2 January 2025
- Published by Elsevier in EBioMedicine
- Vol. 112, 105537
Open AccessGenetic factors shaping the plasma lipidome and the relations to cardiometabolic risk in children and adolescents
- Open Access
- 12 September 2023
- Published by eLife Sciences Publications, Ltd in eLife preprints
- Vol. 12
C. difficile may be overdiagnosed in adults and is a prevalent commensal in infants
Clostridioides difficile is an urgent threat in hospital-acquired infections world-wide, yet the microbial composition associated with C. difficile, in particular in C. difficile infection (CDI) cases, remains poorly characterised. To investigate the gut microbiome composition in CDI patients, we analysed 534 metagenomes from 10 publicly available CDI study populations. We then tracked C. difficile on a global scale, screening 42,900 metagenomes from 253 public studies. Among the CDI cohorts, we detected C. difficile in only 30% of the stool samples from CDI patients. However, we found that multiple other toxigenic species capable of inducing CDI-like symptomatology were prevalent. In addition, the majority of the investigated studies did not adhere to the recommended guidelines for a correct CDI diagnosis. In the global survey, we found that C. difficile prevalence, abundance and biotic context were age-dependent. C. difficile is a rare taxon associated with reduced diversity in healthy adults, but common and associated with increased diversity in infants. We identified a group of species co-occurring with C. difficile exclusively in healthy infants, enriched in obligate anaerobes and in species typical of the healthy adult gut microbiome. C. difficile in healthy infants was therefore associated with multiple indicators of healthy gut microbiome maturation. Our analysis raises concerns about potential CDI overdiagnosis and suggests that C. difficile is an important commensal in infants and that its asymptomatic carriage in adults depends on microbial context.
- 12 September 2023
- Published by eLife Sciences Publications, Ltd in eLife preprints
- Vol. 12
Open AccessC. difficile may be overdiagnosed in adults and is a prevalent commensal in infants
Clostridioides difficile is an urgent threat in hospital-acquired infections world-wide, yet the microbial composition associated with C. difficile, in particular in C. difficile infection (CDI) cases, remains poorly characterised. To investigate the gut microbiome composition in CDI patients, we analysed 534 metagenomes from 10 publicly available CDI study populations. We then tracked C. difficile on a global scale, screening 42,900 metagenomes from 253 public studies. Among the CDI cohorts, we detected C. difficile in only 30% of the stool samples from CDI patients. However, we found that multiple other toxigenic species capable of inducing CDI-like symptomatology were prevalent. In addition, the majority of the investigated studies did not adhere to the recommended guidelines for a correct CDI diagnosis. In the global survey, we found that C. difficile prevalence, abundance and biotic context were age-dependent. C. difficile is a rare taxon associated with reduced diversity in healthy adults, but common and associated with increased diversity in infants. We identified a group of species co-occurring with C. difficile exclusively in healthy infants, enriched in obligate anaerobes and in species typical of the healthy adult gut microbiome. C. difficile in healthy infants was therefore associated with multiple indicators of healthy gut microbiome maturation. Our analysis raises concerns about potential CDI overdiagnosis and suggests that C. difficile is an important commensal in infants and that its asymptomatic carriage in adults depends on microbial context.
- Open Access
- 11 July 2023
- Published by MDPI AG in Diagnostics
- Vol. 13 (14) , 2345
Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
Deep neural networks are complex machine learning models that have shown promising results in analyzing high-dimensional data such as those collected from medical examinations. Such models have the potential to provide fast and accurate medical diagnoses. However, the high complexity makes deep neural networks and their predictions difficult to understand. Providing model explanations can be a way of increasing the understanding of “black box” models and building trust. In this work, we applied transfer learning to develop a deep neural network to predict sex from electrocardiograms. Using the visual explanation method Grad-CAM, heat maps were generated from the model in order to understand how it makes predictions. To evaluate the usefulness of the heat maps and determine if the heat maps identified electrocardiogram features that could be recognized to discriminate sex, medical doctors provided feedback. Based on the feedback, we concluded that, in our setting, this mode of explainable artificial intelligence does not provide meaningful information to medical doctors and is not useful in the clinic. Our results indicate that improved explanation techniques that are tailored to medical data should be developed before deep neural networks can be applied in the clinic for diagnostic purposes.
- 11 July 2023
- Published by MDPI AG in Diagnostics
- Vol. 13 (14) , 2345
Open AccessUsefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
Deep neural networks are complex machine learning models that have shown promising results in analyzing high-dimensional data such as those collected from medical examinations. Such models have the potential to provide fast and accurate medical diagnoses. However, the high complexity makes deep neural networks and their predictions difficult to understand. Providing model explanations can be a way of increasing the understanding of “black box” models and building trust. In this work, we applied transfer learning to develop a deep neural network to predict sex from electrocardiograms. Using the visual explanation method Grad-CAM, heat maps were generated from the model in order to understand how it makes predictions. To evaluate the usefulness of the heat maps and determine if the heat maps identified electrocardiogram features that could be recognized to discriminate sex, medical doctors provided feedback. Based on the feedback, we concluded that, in our setting, this mode of explainable artificial intelligence does not provide meaningful information to medical doctors and is not useful in the clinic. Our results indicate that improved explanation techniques that are tailored to medical data should be developed before deep neural networks can be applied in the clinic for diagnostic purposes.
- Open Access
- 15 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 49772-49784
Multi-Channel Colocalization Analysis and Visualization of Viral Proteins in Fluorescence Microscopy Images
Automatic analysis of colocalizing biological structures in multi-channel fluorescence microscopy images is an important task to quantify and understand biological processes at high spatial-temporal resolution. Here, we introduce a software suite for colocalization analysis of spot-like objects in multi-channel fluorescence microscopy images. The software suite consists of ColocQuant and ColocJ , and is easy to use for biologists. ColocQuant is a Python-based software with graphical user interface to quantify colocalization of particles in two or three channels. Object-based colocalization is performed by an efficient multi-dimensional graph-based k -d-tree approach, which determines nearest neighbors involved in double or triple colocalization. ColocJ enables efficient and intuitive visualization of the color composition of colocalizations by a Maxwell color triangle and a color ribbon. Colocalization information can be visualized for an entire image or a selected region-of-interest. In addition, global statistics of the particle intensity, particle size, and the number of colocalizations over time are provided. The colocalization analysis results can be exported and used in other software. We illustrate the application of our software suite for multi-channel live cell fluorescence microscopy image sequences of viral proteins in hepatitis C virus infected cells. We performed two-channel and three-channel colocalization analysis.
- 15 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 49772-49784
Open AccessMulti-Channel Colocalization Analysis and Visualization of Viral Proteins in Fluorescence Microscopy Images
Automatic analysis of colocalizing biological structures in multi-channel fluorescence microscopy images is an important task to quantify and understand biological processes at high spatial-temporal resolution. Here, we introduce a software suite for colocalization analysis of spot-like objects in multi-channel fluorescence microscopy images. The software suite consists of ColocQuant and ColocJ , and is easy to use for biologists. ColocQuant is a Python-based software with graphical user interface to quantify colocalization of particles in two or three channels. Object-based colocalization is performed by an efficient multi-dimensional graph-based k -d-tree approach, which determines nearest neighbors involved in double or triple colocalization. ColocJ enables efficient and intuitive visualization of the color composition of colocalizations by a Maxwell color triangle and a color ribbon. Colocalization information can be visualized for an entire image or a selected region-of-interest. In addition, global statistics of the particle intensity, particle size, and the number of colocalizations over time are provided. The colocalization analysis results can be exported and used in other software. We illustrate the application of our software suite for multi-channel live cell fluorescence microscopy image sequences of viral proteins in hepatitis C virus infected cells. We performed two-channel and three-channel colocalization analysis.