References of "Glaab, Enrico 50001863"
     in
Bookmark and Share    
Full Text
Peer Reviewed
See detailBiomedical and Clinical Research Data Management
Ganzinger, Matthias; Glaab, Enrico UL; Kerssemakers, Jules et al

in Wolkenhauer, Olaf (Ed.) Systems Medicine - Integrative, Qualitative and Computational Approaches (in press)

Detailed reference viewed: 72 (3 UL)
Full Text
Peer Reviewed
See detailA novel TIMP3 mutation associated with a retinitis pigmentosa-like phenotype
DeBenedictis, Meghan; Gindzin, Yosef; Glaab, Enrico UL et al

in Ophthalmic Genetics (in press)

Sorsby Fundus Dystrophy is an inherited macular degeneration caused by pathogenic variants in the TIMP3 gene. In this study we describe a father and son initially diagnosed with retinitis pigmentosa of ... [more ▼]

Sorsby Fundus Dystrophy is an inherited macular degeneration caused by pathogenic variants in the TIMP3 gene. In this study we describe a father and son initially diagnosed with retinitis pigmentosa of unknown genetic origin. More recent genetic testing of the patients, identified a novel c.410A>G; p.Tyr137Cys variant of uncertain clinical significance in the Tissue Inhibitor of Metalloproteinase-3 (TIMP3) gene. The atypical clinical findings led us to compare the theoretical molecular effects of this variant on the TIMP3 protein structure and interactions with other proteins using homology modeling and machine learning predictions. [less ▲]

Detailed reference viewed: 111 (3 UL)
Full Text
Peer Reviewed
See detailPredictive and interpretable models via the stacked elastic net
Rauschenberger, Armin UL; Glaab, Enrico UL; van de Wiel, Mark

in Bioinformatics (in press)

Motivation: Machine learning in the biomedical sciences should ideally provide predictive and interpretable models. When predicting outcomes from clinical or molecular features, applied researchers often ... [more ▼]

Motivation: Machine learning in the biomedical sciences should ideally provide predictive and interpretable models. When predicting outcomes from clinical or molecular features, applied researchers often want to know which features have effects, whether these effects are positive or negative, and how strong these effects are. Regression analysis includes this information in the coefficients but typically renders less predictive models than more advanced machine learning techniques. Results: Here we propose an interpretable meta-learning approach for high-dimensional regression. The elastic net provides a compromise between estimating weak effects for many features and strong effects for some features. It has a mixing parameter to weight between ridge and lasso regularisation. Instead of selecting one weighting by tuning, we combine multiple weightings by stacking. We do this in a way that increases predictivity without sacrificing interpretability. Availability and Implementation: The R package starnet is available on GitHub: https://github.com/rauschenberger/starnet. [less ▲]

Detailed reference viewed: 166 (12 UL)
Full Text
Peer Reviewed
See detailGBA variants in Parkinson’s disease: clinical, metabolomic and multimodal neuroimaging phenotypes
Greuel, Andrea; Trezzi, Jean-Pierre UL; Glaab, Enrico UL et al

in Movement Disorders (in press)

Background: Alterations in the GBA gene (NM_000157.3) are the most important genetic risk factor for Parkinson’s disease. Biallelic GBA mutations cause the lysosomal storage disorder Gaucher´s disease ... [more ▼]

Background: Alterations in the GBA gene (NM_000157.3) are the most important genetic risk factor for Parkinson’s disease. Biallelic GBA mutations cause the lysosomal storage disorder Gaucher´s disease. The GBA variants p.E365K and p.T408M are associated with Parkinson’s but not with Gaucher´s disease. The pathophysiological role of these variants needs to be further explored. Objective: This study analyzed the clinical, neuropsychological, metabolic and neuroimaging phenotypes of Parkinson’s disease patients carrying the GBA variants p.E365K and p.T408M. Methods: GBA was sequenced in 56 mid-stage Parkinson’s disease patients. Carriers of GBA variants were compared to non-carriers regarding clinical history and symptoms, neuropsychological features, metabolomics and multimodal neuroimaging. Blood plasma gas chromatography coupled to mass spectrometry, [18F]FDopa PET, [18F]FDG PET, and resting-state fMRI were performed. Results: Sequence analysis detected 13 heterozygous GBA variant carriers (seven with p.E365K, six with p.T408M). One patient carried a GBA mutation (p.N409S) and was excluded. Clinical history and symptoms were not significantly different between groups. Global cognitive performance was lower in variant carriers. Metabolomic group differences were suggestive of more severe Parkinson’s disease-related alterations in carriers versus non-carriers. [18F]FDopa and [18F]FDG PET showed signs of a more advanced disease; [18F]FDG PET and fMRI showed similarities with Lewy body dementia and Parkinson’s disease dementia in carriers. Conclusions: This is the first study to comprehensively assess (neuro-)biological phenotypes of GBA variants in Parkinson’s disease. Metabolomics and neuroimaging detected more significant group differences than clinical and behavioral evaluation. These alterations could be promising to monitor effects of disease-modifying treatments targeting glucocerebrosidase metabolism. [less ▲]

Detailed reference viewed: 88 (3 UL)
Full Text
Peer Reviewed
See detailA new synuclein-transgenic mouse model for early Parkinson's reveals molecular features of preclinical disease
Hendrickx, Diana M.; Garcia, Pierre; Ashrafi, Amer et al

in Molecular Neurobiology (in press)

Understanding Parkinson’s disease (PD) in particular in its earliest phases, is important for diagnosis and treatment. However, human brain samples are collected post- mortem, reflecting mainly end stage ... [more ▼]

Understanding Parkinson’s disease (PD) in particular in its earliest phases, is important for diagnosis and treatment. However, human brain samples are collected post- mortem, reflecting mainly end stage disease. Because brain samples of mouse models can be collected at any stage of the disease process, they are useful to investigate PD progression. Here, we compare ventral midbrain transcriptomics profiles from α- synuclein transgenic mice with a progressive, early PD-like striatal neurodegeneration across different ages using pathway, gene set and network analysis methods. Our study uncovers statistically significant altered genes across ages and between genotypes with known, suspected, or unknown function in PD pathogenesis and key pathways associated with disease progression. Among those are genotype-dependent alterations associated with synaptic plasticity, neurotransmission, as well as mitochondria-related genes and dysregulation of lipid metabolism. Age-dependent changes were among others observed in neuronal and synaptic activity, calcium homeostasis, and membrane receptor signaling pathways, many of which linked to G- protein coupled receptors. Most importantly, most changes occurred before neurodegeneration was detected in this model, which points to a sequence of gene expression events that may be relevant for disease initiation and progression. It is tempting to speculate that molecular changes similar to those changes observed in our model happen in midbrain dopaminergic neurons before they start to degenerate. In other words, we believe we have uncovered molecular changes that accompany the progression from preclinical to early PD. [less ▲]

Detailed reference viewed: 103 (0 UL)
Full Text
Peer Reviewed
See detailGene selection for optimal prediction of cell position in tissues from single-cell transcriptomics
Tanevski, Jovan; Nguyen, Thin; Truong, Buu et al

in Life Science Alliance (in press)

Single-cell RNA-seq (scRNAseq) technologies are rapidly evolving. While very informative, in standard scRNAseq experiments the spatial organization of the cells in the tissue of origin is lost. Conversely ... [more ▼]

Single-cell RNA-seq (scRNAseq) technologies are rapidly evolving. While very informative, in standard scRNAseq experiments the spatial organization of the cells in the tissue of origin is lost. Conversely, spatial RNA-seq technologies designed to maintain cell localization have limited throughput and gene coverage. Mapping scRNAseq to genes with spatial information increases coverage while providing spatial location. However, methods to perform such mapping have not yet been benchmarked. To fill this gap, we organized the DREAM Single-Cell Transcriptomics challenge focused on the spatial reconstruction of cells from the Drosophila embryo from scRNAseq data, leveraging as silver standard, genes with in situ hybridization data from the Berkeley Drosophila Transcription Network Project reference atlas. The 34 participating teams used diverse algorithms for gene selection and location prediction, while being able to correctly localize clusters of cells. Selection of predictor genes was essential for this task. Predictor genes showed a relatively high expression entropy, high spatial clustering and included prominent developmental genes such as gap and pair-rule genes and tissue markers. Application of the Top-10 methods to a zebrafish embryo dataset yielded similar performance and statistical properties of the selected genes than in the Drosophila data. This suggests that methods developed in this challenge are able to extract generalizable properties of genes that are useful to accurately reconstruct the spatial arrangement of cells in tissues. [less ▲]

Detailed reference viewed: 116 (3 UL)
Full Text
Peer Reviewed
See detailNon-Coding RNAs in the Brain-Heart Axis: The Case of Parkinson’s Disease
Acharya, Shubhra; Salgado-Somoza, Antonio; Stefanizzi, Francesca Maria et al

in International Journal of Molecular Sciences (2020), 21(18), 6513

Parkinson’s disease (PD) is a complex and heterogeneous disorder involving multiple genetic and environmental influences. Although a wide range of PD risk factors and clinical markers for the symptomatic ... [more ▼]

Parkinson’s disease (PD) is a complex and heterogeneous disorder involving multiple genetic and environmental influences. Although a wide range of PD risk factors and clinical markers for the symptomatic motor stage of the disease have been identified, there are still no reliable biomarkers available for the early pre-motor phase of PD and for predicting disease progression. High-throughput RNA-based biomarker profiling and modeling may provide a means to exploit the joint information content from a multitude of markers to derive diagnostic and prognostic signatures. In the field of PD biomarker research, currently, no clinically validated RNA-based biomarker models are available, but previous studies reported several significantly disease-associated changes in RNA abundances and activities in multiple human tissues and body fluids. Here, we review the current knowledge of the regulation and function of non-coding RNAs in PD, focusing on microRNAs, long non-coding RNAs, and circular RNAs. Since there is growing evidence for functional interactions between the heart and the brain, we discuss the benefits of studying the role of non-coding RNAs in organ interactions when deciphering the complex regulatory networks involved in PD progression. We finally review important concepts of harmonization and curation of high throughput datasets, and we discuss the potential of systems biomedicine to derive and evaluate RNA biomarker signatures from high-throughput expression data. [less ▲]

Detailed reference viewed: 72 (1 UL)
Full Text
Peer Reviewed
See detailComparative transcriptome analysis of Parkinson’s disease and Hutchinson-Gilford progeria syndrome reveals shared susceptible cellular network processes
Hendrickx, Diana M.; Glaab, Enrico UL

in BMC Medical Genomics (2020), 13(114),

Background Parkinson’s Disease (PD) and Hutchinson-Gilford Progeria Syndrome (HGPS) are two heterogeneous disorders, which both display molecular and clinical alterations associated with the aging process ... [more ▼]

Background Parkinson’s Disease (PD) and Hutchinson-Gilford Progeria Syndrome (HGPS) are two heterogeneous disorders, which both display molecular and clinical alterations associated with the aging process. However, similarities and differences between molecular changes in these two disorders have not yet been investigated systematically at the level of individual biomolecules and shared molecular network alterations. Methods Here, we perform a comparative meta-analysis and network analysis of human transcriptomics data from case-control studies for both diseases to investigate common susceptibility genes and sub-networks in PD and HGPS. Alzheimer’s disease (AD) and primary melanoma (PM) were included as controls to confirm that the identified overlapping susceptibility genes for PD and HGPS are non-generic. Results We find statistically significant, overlapping genes and cellular processes with significant alterations in both diseases. Interestingly, the majority of these shared affected genes display changes with opposite directionality, indicating that shared susceptible cellular processes undergo different mechanistic changes in PD and HGPS. A complementary regulatory network analysis also reveals that the altered genes in PD and HGPS both contain targets controlled by the upstream regulator CDC5L. Conclusions Overall, our analyses reveal a significant overlap of affected cellular processes and molecular sub-networks in PD and HGPS, including changes in aging-related processes that may reflect key susceptibility factors associated with age-related risk for PD. [less ▲]

Detailed reference viewed: 82 (0 UL)
Full Text
Peer Reviewed
See detailVariants in Miro1 cause alterations of ER-mitochondria contact sites in fibroblasts from Parkinson's disease patients
Berenguer, Clara UL; Grossmann, Dajana; Massart, François UL et al

in Journal of Clinical Medicine (2019)

Background: Although most cases of Parkinson´s disease (PD) are idiopathic with unknown cause, an increasing number of genes and genetic risk factors have been discovered that play a role in PD ... [more ▼]

Background: Although most cases of Parkinson´s disease (PD) are idiopathic with unknown cause, an increasing number of genes and genetic risk factors have been discovered that play a role in PD pathogenesis. Many of the PD‐associated proteins are involved in mitochondrial quality control, e.g., PINK1, Parkin, and LRRK2, which were recently identified as regulators of mitochondrial‐endoplasmic reticulum (ER) contact sites (MERCs) linking mitochondrial homeostasis to intracellular calcium handling. In this context, Miro1 is increasingly recognized to play a role in PD pathology. Recently, we identified the first PD patients carrying mutations in RHOT1, the gene coding for Miro1. Here, we describe two novel RHOT1 mutations identified in two PD patients and the characterization of the cellular phenotypes. Methods: Using whole exome sequencing we identified two PD patients carrying heterozygous mutations leading to the amino acid exchanges T351A and T610A in Miro1. We analyzed calcium homeostasis and MERCs in detail by live cell imaging and immunocytochemistry in patient‐derived fibroblasts. Results: We show that fibroblasts expressing mutant T351A or T610A Miro1 display impaired calcium homeostasis and a reduced amount of MERCs. All fibroblast lines from patients with pathogenic variants in Miro1, revealed alterations of the structure of MERCs. Conclusion: Our data suggest that Miro1 is important for the regulation of the structure and function of MERCs. Moreover, our study supports the role of MERCs in the pathogenesis of PD and further establishes variants in RHOT1 as rare genetic risk factors for neurodegeneration. [less ▲]

Detailed reference viewed: 81 (9 UL)
Full Text
See detailComputational analysis of molecular network perturbations in complex diseases
Glaab, Enrico UL

Presentation (2019, November 01)

Detailed reference viewed: 142 (5 UL)
Full Text
Peer Reviewed
See detailMutations in RHOT1 disrupt ER-mitochondria contact sites interfering with calcium homeostasis and mitochondrial dynamics in Parkinson's disease.
Grossmann, Dajana UL; Berenguer, Clara UL; Bellet, Marie Estelle et al

in Antioxidants & redox signaling (2019)

OBJECTIVE: The outer mitochondrial membrane protein Miro1 is a crucial player in mitochondrial dynamics and calcium homeostasis. Recent evidence indicated that Miro1 mediates calcium-induced mitochondrial ... [more ▼]

OBJECTIVE: The outer mitochondrial membrane protein Miro1 is a crucial player in mitochondrial dynamics and calcium homeostasis. Recent evidence indicated that Miro1 mediates calcium-induced mitochondrial shape transition (MiST), which is a prerequisite for the initiation of mitophagy. Moreover, altered Miro1 protein levels have emerged as a shared feature of monogenic and sporadic Parkinson's disease (PD), but, so far, no disease-associated variants in RHOT1 have been identified. RESULTS: Here, for the first time, we describe heterozygous RHOT1 mutations in two PD patients (het c.815G>A; het c.1348C>T) and identified mitochondrial phenotypes with reduced mitochondrial mass in patient-derived cellular models. Both mutations lead to decreased ER-mitochondrial contact sites and calcium dyshomeostasis. As a consequence, energy metabolism was impaired, which in turn lead to increased mitophagy. CONCLUSION: In summary, our data support the role of Miro1 in maintaining calcium homeostasis and mitochondrial quality control in PD. [less ▲]

Detailed reference viewed: 274 (32 UL)
Full Text
Peer Reviewed
See detailTranscriptome profiling data reveals Ubiquitin-Specific Peptidase 9 knockdown effects
Glaab, Enrico UL; Antony, Paul UL; Köglsberger, Sandra et al

in Data in Brief (2019), 25(1), 104130

Ubiquitin specific peptidase 9 (USP9) is a deubiquitinase encoded by a sex-linked gene with a Y-chromosomal form (USP9Y) and an X-chromosomal form (USP9X) that escapes X-inactivation. Since USP9 is a key ... [more ▼]

Ubiquitin specific peptidase 9 (USP9) is a deubiquitinase encoded by a sex-linked gene with a Y-chromosomal form (USP9Y) and an X-chromosomal form (USP9X) that escapes X-inactivation. Since USP9 is a key regulatory gene with sex-linked expression in the human brain, the gene may be of interest for researchers studying molecular gender differences and ubiquitin signaling in the brain. To assess the downstream effects of knocking down USP9X and USP9Y on a transcriptome-wide scale, we have conducted microarray profiling experiments using the human DU145 prostate cancer cell culture model, after confirming the robust expression of both USP9X and USP9Y in this model. By designing shRNA constructs for the specific knockdown of USP9X and the joint knockdown of USP9X and USP9Y, we have compared gene expression changes in both knockdowns to control conditions to infer potential shared and X- or Y-form specific alterations. Here, we provide details of the corresponding microarray profiling data, which has been deposited in the Gene Expression Omnibus database (GEO series accession number GSE79376). A biological interpretation of the data in the context of a potential involvement of USP9 in Alzheimer’s disease has previously been presented in Köglsberger et al. (2016). To facilitate the re-use and re-analysis of the data for other applications, e.g. the study of ubiquitin signaling and protein turnover control, and the regulation of molecular gender differences in the human brain and brain-related disorders, we provide a more in-depth discussion of the data properties, specifications and possible use cases. [less ▲]

Detailed reference viewed: 120 (3 UL)
Full Text
Peer Reviewed
See detailBSA4Yeast: Web-based quantitative trait locus linkage analysis and bulk segregant analysis of yeast sequencing data
Zhang, Zhi; Jung, Paul; Groues, Valentin UL et al

in GigaScience (2019), 8(6), 060

Quantitative Trait Loci (QTL) mapping using bulk segregants is an effective approach for identifying genetic variants associated with phenotypes of interest in model organisms. By exploiting next ... [more ▼]

Quantitative Trait Loci (QTL) mapping using bulk segregants is an effective approach for identifying genetic variants associated with phenotypes of interest in model organisms. By exploiting next-generation sequencing technology, the QTL mapping accuracy can be improved significantly, providing a valuable means to annotate new genetic variants. However, setting up a comprehensive analysis framework for this purpose is a time-consuming and error prone task, posing many challenges for scientists with limited experience in this domain. Findings: Here, we present BSA4Yeast, a comprehensive web-application for QTL mapping via bulk segregant analysis of yeast sequencing data. The software provides an automated and efficiency-optimized data processing, up-to-date functional annotations, and an interactive web-interface to explore identified QTLs. Conclusion: BSA4Yeast enables researchers to identify plausible candidate genes in QTL regions efficiently in order to validate their genetic variations experimentally as causative for a phenotype of interest. BSA4Yeast is freely available at https://bsa4yeast.lcsb.uni.lu. [less ▲]

Detailed reference viewed: 207 (18 UL)
Full Text
Peer Reviewed
See detailIntegrative analysis of blood metabolomics and PET brain neuroimaging data for Parkinson's disease
Glaab, Enrico UL; Trezzi, Jean-Pierre UL; Greuel, Andrea et al

in Neurobiology of Disease (2019), 124(1), 555-562

The diagnosis of Parkinson's disease (PD) often remains a clinical challenge. Molecular neuroimaging can facilitate the diagnostic process. The diagnostic potential of metabolomic signatures has recently ... [more ▼]

The diagnosis of Parkinson's disease (PD) often remains a clinical challenge. Molecular neuroimaging can facilitate the diagnostic process. The diagnostic potential of metabolomic signatures has recently been recognized. Methods: We investigated whether the joint data analysis of blood metabolomics and PET imaging by machine learning provides enhanced diagnostic discrimination and gives further pathophysiological insights. Blood plasma samples were collected from 60 PD patients and 15 age- and gender-matched healthy controls. We determined metabolomic profiles by gas chromatography coupled to mass spectrometry (GC-MS). In the same cohort and at the same time we performed FDOPA PET in 44 patients and 14 controls and FDG PET in 51 patients and 16 controls. 18 PD patients were available for a follow-up exam after one year. Both data sets were analysed by two machine learning approaches, applying either linear support vector machines or random forests within a leave-one-out cross-validation and computing receiver operating characteristic (ROC) curves. Results: In the metabolomics data, the baseline comparison between cases and controls as well as the followup assessment of patients pointed to metabolite changes associated with oxidative stress and inflammation. For the FDOPA and FDG PET data, the diagnostic predictive performance (DPP) in the ROC analyses was highest when combining imaging features with metabolomics data (ROC AUC for best FDOPA + metabolomics model: 0.98; AUC for best FDG + metabolomics model: 0.91). DPP was lower when using only PET attributes or only metabolomics signatures. Conclusion: Integrating blood metabolomics data combined with PET data considerably enhances the diagnostic discrimination power. Metabolomic signatures also indicate interesting disease-inherent changes in cellular processes, including oxidative stress response and inflammation. [less ▲]

Detailed reference viewed: 149 (26 UL)
Full Text
Peer Reviewed
See detailImpaired serine metabolism complements LRRK2-G2019S pathogenicity in PD patients
Nickels, Sarah UL; Walter, Jonas; Bolognin, Silvia UL et al

in Parkinsonism and Related Disorders (2019)

Detailed reference viewed: 175 (26 UL)
Full Text
Peer Reviewed
See detailIntegrated Analyses of Microbiome and Longitudinal Metabolome Data Reveal Microbial-Host Interactions on Sulfur Metabolism in Parkinson’s Disease
Hertel, Johannes; Harms, Amy C.; Heinken, Almut et al

in Cell Reports (2019), 29(7), 1767-1777

Parkinson’s disease (PD) exhibits systemic effects on human metabolism with emerging roles for the gut microbiome. Here, we integrated longitudinal metabolome data from 30 drug-naïve, de-novo PD patients ... [more ▼]

Parkinson’s disease (PD) exhibits systemic effects on human metabolism with emerging roles for the gut microbiome. Here, we integrated longitudinal metabolome data from 30 drug-naïve, de-novo PD patients and 30 matched controls with constraint-based modeling of gut microbial communities derived from an independent, drug-naïve PD cohort, and prospective data from a general population. Our key results are i) longitudinal trajectory of metabolites associated with the interconversion of methionine and cysteine via cystathionine differed between PD patients and controls, ii) dopaminergic medication showed strong lipidomic signatures, iii) taurine-conjugated bile acids correlated with the severity of motor symptoms, while low levels of sulfated taurolithocholate were associated with incident PD in the general population, and iv) computational modeling predicted changes in sulfur metabolism, driven by A. muciniphila and B. wadsworthia, consistent with the changed metabolome. In conclusion, the multi-omics integration revealed PD-specific patterns in microbial-host sulfur co-metabolism that may contribute to PD severity. [less ▲]

Detailed reference viewed: 120 (16 UL)
Full Text
Peer Reviewed
See detail3D Cultures of Parkinson's Disease‐Specific Dopaminergic Neurons for High Content Phenotyping and Drug Testing
Bolognin, Silvia UL; Fossépré, Marie; Qing, Xiaobing et al

in Advanced Science (2018)

Parkinson's disease (PD)‐specific neurons, grown in standard 2D cultures, typically only display weak endophenotypes. The cultivation of PD patient‐specific neurons, derived from induced pluripotent stem ... [more ▼]

Parkinson's disease (PD)‐specific neurons, grown in standard 2D cultures, typically only display weak endophenotypes. The cultivation of PD patient‐specific neurons, derived from induced pluripotent stem cells carrying the LRRK2‐G2019S mutation, is optimized in 3D microfluidics. The automated image analysis algorithms are implemented to enable pharmacophenomics in disease‐relevant conditions. In contrast to 2D cultures, this 3D approach reveals robust endophenotypes. High‐content imaging data show decreased dopaminergic differentiation and branching complexity, altered mitochondrial morphology, and increased cell death in LRRK2‐G2019S neurons compared to isogenic lines without using stressor agents. Treatment with the LRRK2 inhibitor 2 (Inh2) rescues LRRK2‐G2019S‐dependent dopaminergic phenotypes. Strikingly, a holistic analysis of all studied features shows that the genetic background of the PD patients, and not the LRRK2‐G2019S mutation, constitutes the strongest contribution to the phenotypes. These data support the use of advanced in vitro models for future patient stratification and personalized drug development. [less ▲]

Detailed reference viewed: 290 (38 UL)
Full Text
Peer Reviewed
See detailCombining PET imaging and blood metabolomics data to improve machine learning models for Parkinson’s disease diagnosis
Glaab, Enrico UL; Trezzi, Jean-Pierre UL; Greuel et al

Poster (2018, October 08)

Objective: To investigate whether the integration of PET imaging and metabolomics data can provide improved machine learning models for PD diagnosis. Background: The reliable diagnosis of PD can remain ... [more ▼]

Objective: To investigate whether the integration of PET imaging and metabolomics data can provide improved machine learning models for PD diagnosis. Background: The reliable diagnosis of PD can remain challenging, even at the motor stage. PET imaging can be used to confirm the clinical diagnosis. However, limitations in the robustness of predictive features extracted from the data and the costs associated with PET imaging restrict its application. Using blood metabolomics data as an additional information source may provide improved combined diagnostic models and/or an initial filter to decide on whether to apply PET imaging. Methods: Metabolomics profiling of blood plasma samples using gas chromatography coupled to mass spectrometry (GC­MS) was conducted in 60 IPD patients and 15 healthy controls. After pre-processing, these data were compared to neuroimaging data for subsets of the same individuals using FDOPA PET (44 patients and 14 controls) and FDG PET (51 patients and 15 controls). Machine learning models using linear support vector machines were trained on 50% of the data and evaluated on a 50% hold­out test set using Receiver Operating Characteristic (ROC) curves. Next, standardized FDOPA and FDG PET intensity measurements were combined with those from the metabolomics data to build and evaluate sample classification models in the same manner as for the individual datasets. Results: Both for the FDOPA and FDG PET data, the predictive performance given by the area under the ROC curve (AUC) was highest when combining imaging features with those from the metabolomics data (AUC for FDOPA + metabolomics: 0.98; AUC for FDG + metabolomics: 0.91). The performance was generally lower when using only the respective PET attributes (FDOPA: 0.94, FDG: 0.8) or only the metabolomics data (AUC: 0.66). [less ▲]

Detailed reference viewed: 151 (21 UL)