Improving Machine Learning-based Prediction of Frailty in Elderly People with Digital Wearables : Data from the Berlin Aging Study II (BASE-II)

Poster (2022, October 26)

Frailty is a geriatric medical condition that is highly associated with age and age-related diseases. The multidimensional consequences of frailty are heavily impacting the quality of life, and will ... [more ▼]

Frailty is a geriatric medical condition that is highly associated with age and age-related diseases. The multidimensional consequences of frailty are heavily impacting the quality of life, and will inevitably increase the burden on healthcare systems in the future. Most importantly, the lack of a universal standard to describe, diagnose, or let alone treat frailty, is further complicating the situation in the long-term. Nowadays, more and more frailty assessment tools are being developed on a regional and institutional basis, which is continuing to drive the heterogeneity in the characterization of frailty further apart. Gaining better insights into the underlying causes and pathophysiology of frailty, and how it is developing in patients is, therefore, required to establish strong and accurately tailored response schemes for frail patients, where currently only symptoms are treated. Thus, in this study, we deployed machine learning-based classification and optimization techniques to predict frailty in elderly people aged 65 or above from the Berlin Aging Study II (BASE-II, n=1512, frail=484) and revealed some of the most informative biomedical information to characterize frailty, including new potential biomarkers. Frailty in BASE-II was measured by the Fried et al. 5-item frailty index, composed of the clinical variables grip strength, weight loss, exhaustion, physical activity, and gait. The level of frailty in BASE-II was adapted for binary classification purposes by merging the pre-frail and frail levels as frail. A configurable in-house pipeline was developed for pre-processing the clinical data and predicting the target disease by deploying Support Vector Machines Classification. The most informative and essential subgroup of clinical measurements with regards to frailty was investigated by re-optimizing an initially full data-driven model by sequentially leaving out one subgroup. The best prediction power was yielded with resampling and dimensionality reduction techniques using the F-beta-2 score, and was further improved by adding one item of the Fried et al. frailty index. Furthermore, differences between the gender in the data set led to the investigation of gender-specific model configurations, followed by re-optimizations. As a result, we were able to specifically increase the predictive power in gender-specific groups, and will simultaneously emphasize on the differences between the most informative clinical biomarkers as well as the most essential subgroups for mixed and gender-specific BASE-II. The results herein suggest that a combination of the detected easy-to-obtain biomedical information on frailty risk factors together with one Fried et al. phenotype information provided by i.e., smart wearable devices (gait, grip strength, …) could significantly improve the frailty prediction power in mixed and gender-specific clinical cohort data. [less ▲]

SnapShot: The Expobiome Map

in Cell Host and Microbe (2022), 30(9), 1340

The human gut microbiome is intricately connected to health and disease. Microbiome-derived molecules are implicated in many chronic diseases involving inflammation. Herein, we summarize the diverse ... [more ▼]

The human gut microbiome is intricately connected to health and disease. Microbiome-derived molecules are implicated in many chronic diseases involving inflammation. Herein, we summarize the diverse complex of such immunogenic molecules, including nucleic acids, (poly)peptides, structural molecules, and metabolites. The interactions between this “expobiome” and human immune pathways are specifically illustrated in the context of chronic diseases. [less ▲]

An archaeal compound as a driver of Parkinson’s disease pathogenesis

E-print/Working paper (2022)

Patients with Parkinson’s disease (PD) exhibit differences in their gut microbiomes compared to healthy individuals. Although differences have most commonly been described in the abundances of bacterial ... [more ▼]

Patients with Parkinson’s disease (PD) exhibit differences in their gut microbiomes compared to healthy individuals. Although differences have most commonly been described in the abundances of bacterial taxa, changes to viral and archaeal populations have also been observed. Mechanistic links between gut microbes and PD pathogenesis remain elusive but could involve molecules that promote α-synuclein aggregation. Here, we show that 2-hydroxypyridine (2-HP) represents a key molecule for the pathogenesis of PD. We observe significantly elevated 2-HP levels in faecal samples from patients with PD or its prodrome, idiopathic REM sleep behaviour disorder (iRBD), compared to healthy controls. 2-HP is correlated with the archaeal species Methanobrevibacter smithii and with genes involved in methane metabolism, and it is detectable in isolate cultures of M. smithii. We demonstrate that 2-HP is selectively toxic to transgenic α-synuclein overexpressing yeast and increases α-synuclein aggregation in a yeast model as well as in human induced pluripotent stem cell derived enteric neurons. It also exacerbates PD-related motor symptoms, α-synuclein aggregation, and striatal degeneration when injected intrastriatally in transgenic mice overexpressing human α-synuclein. Our results highlight the effect of an archaeal molecule in relation to the gut-brain axis, which is critical for the diagnosis, prognosis, and treatment of PD. [less ▲]

Evaluation of the Molecular Pathogenesis of Adrenocortical Tumors by Whole-Genome Sequencing

in Journal of the Endocrine Society (2021, May 03), 5(Issue Supplement_1), 68

Pathogenesis of autonomous steroid secretion and adrenocortical tumorigenesis remains partially obscure. Our aim was to identify novel genetic alterations in adrenocortical adenomas (ACA) without somatic ... [more ▼]

Pathogenesis of autonomous steroid secretion and adrenocortical tumorigenesis remains partially obscure. Our aim was to identify novel genetic alterations in adrenocortical adenomas (ACA) without somatic mutations in known driver genes. Whole-genome sequencing was performed on 26 ACA/blood-derived DNA pairs without driver mutations in PRKACA, GNAS and CTNNB1 genes at previous WES (ENSAT study JCEM 2016). These included 12 cortisol-producing adenomas with Cushing syndrome (CS-CPAs), 7 with mild autonomous cortisol secretion (MACS-CPAs), and 7 endocrine-inactive ACAs (EIAs). Seven adrenocortical carcinomas (ACC) were added to the cohort. We developed a bioinformatics pipeline for a comprehensive genome analysis and to reveal differences in variant distribution. Strelka, VarScan2 and ANNOVAR software and an in-house confidence score were used for variant calling and functional annotation. Combined Annotation-Dependent-Depletion (CADD) values were used to prioritize pathogenic variants. Additional focus relied on variants in pathogenically known pathways (Wnt/β-catenin, cAMP/PKA pathway). NovoBreak algorithm was applied to discover structural variations. Two hypermutated CS-CPA samples were excluded from further analysis. Using different filters, we detected variants in driver genes not observed at WES (one p.S45P in CTNNB1 and one p.R206L in PRKACA in two different CS-CPAs). In total, we report 179,830 variations (179,598 SNVs; 232 indels) throughout all samples, being more abundant in ACC (88,954) compared to ACA (CS-CPAs: 31,821; MACS-CPAs: 35,008; EIAs: 29,963). Most alterations were in intergenic (>50%), followed by intronic and ncRNA intronic regions. A total of 32 predicted pathogenic variants were found in both coding (CADD values ≥ 15) and non-coding (CADD values ≥ 5) regions. We found 3,301 possibly damaging and recurrent variants (intergenic mutations removed) (CS-CPAs: 1,463; MACS-CPAs: 1,549; EIAs: 1,268; ACC: 1,660), mostly accumulated in intronic regions. Some of these were detected in members of the Wnt/β-catenin (CS-CPAs: 6; MACS-CPAs: 2; EIA: 1) and cAMP/PKA (CS-CPAs: 6; MACS-CPAs: 7; EIA: 4) pathways (e.g. ADCY1, ADCY2, GNA13, PDE11A). We also found a slightly higher number of structural variations in EIA (3,620) and ACC (3,486) compared to CS-CPAs (977) and MACS-CPAs (2,119). In conclusion, still unrevealed genetic alterations, especially in intronic regions, may accompany early adrenal tumorigenesis and/or autonomous cortisol secretion. [less ▲]

Gene expression profile of CD14+ blood monocytes following lifestyle-induced weight loss in individuals with metabolic syndrome.

in Scientific Reports (2020), 10(1), 17855

Lifestyle-induced weight loss is regarded as an efficient therapy to reverse metabolic syndrome (MetS) and to prevent disease progression. The objective of this study was to investigate whether lifestyle ... [more ▼]

Lifestyle-induced weight loss is regarded as an efficient therapy to reverse metabolic syndrome (MetS) and to prevent disease progression. The objective of this study was to investigate whether lifestyle-induced weight loss modulates gene expression in circulating monocytes. We analyzed and compared gene expression in monocytes (CD14+ cells) and subcutaneous adipose tissue biopsies by unbiased mRNA profiling. Samples were obtained before and after diet-induced weight loss in well-defined male individuals in a prospective controlled clinical trial (ICTRP Trial Number: U1111-1158-3672). The BMI declined significantly (- 12.6%) in the treatment arm (N = 39) during the 6-month weight loss intervention. This was associated with a significant reduction in hsCRP (- 45.84%) and circulating CD14+ cells (- 21.0%). Four genes were differentially expressed (DEG's) in CD14+ cells following weight loss (ZRANB1, RNF25, RB1CC1 and KMT2C). Comparative analyses of paired CD14+ monocytes and subcutaneous adipose tissue samples before and after weight loss did not identify common genes differentially regulated in both sample types. Lifestyle-induced weight loss is associated with specific changes in gene expression in circulating CD14+ monocytes, which may affect ubiquitination, histone methylation and autophagy [less ▲]

Measuring mode I fracture properties of thick-layered structural silicone sealants

in International Journal of Adhesion and Adhesives (2019), 91

The failure hyperelastic structural sealant joints is a fracture mechanics problem. Its modeling requires knowledge of the material property fracture toughness. The present work describes a methodology ... [more ▼]

The failure hyperelastic structural sealant joints is a fracture mechanics problem. Its modeling requires knowledge of the material property fracture toughness. The present work describes a methodology for determining the mode I bulk material fracture toughness of structural silicone sealants. The concept is demonstrated using DOWSIL™ 993 as an example. In total we manufactured, tested and evaluated 13 DCB specimens of 6mm and 12mm adhesive thickness. The fracture toughness is identified from J-integral measurements. The theory for determining the J-integral at finite deformations is laid out and an automated data analysis procedure is suggested. The presented approach further allows... [less ▲]

Rare ABCA7 variants in 2 German families with Alzheimer disease

in Neurology Genetics (2018), 4(2),

Objective The aim of this study was to identify variants associated with familial late-onset Alzheimer disease (AD) using whole-genome sequencing. Methods Several families with an autosomal dominant ... [more ▼]

Objective The aim of this study was to identify variants associated with familial late-onset Alzheimer disease (AD) using whole-genome sequencing. Methods Several families with an autosomal dominant inheritance pattern of AD were analyzed by whole-genome sequencing. Variants were prioritized for rare, likely pathogenic variants in genes already known to be associated with AD and confirmed by Sanger sequencing using standard protocols. Results We identified 2 rare ABCA7 variants (rs143718918 and rs538591288) with varying penetrance in 2 independent German AD families, respectively. The single nucleotide variant (SNV) rs143718918 causes a missense mutation, and the deletion rs538591288 causes a frameshift mutation of ABCA7. Both variants have previously been reported in larger cohorts but with incomplete segregation information. ABCA7 is one of more than 20 AD risk loci that have so far been identified by genome-wide association studies, and both common and rare variants of ABCA7 have previously been described in different populations with higher frequencies in AD cases than in controls and varying penetrance. Furthermore, ABCA7 is known to be involved in several AD-relevant pathways. Conclusions We conclude that both SNVs might contribute to the development of AD in the examined family members. Together with previous findings, our data confirm ABCA7 as one of the most relevant AD risk genes. [less ▲]

The Gut Microbiota and Hematopoietic Stem Cell Transplantation: Challenges and Potentials.

in Journal of innate immunity (2018)

The human gut microbiota gained tremendous importance in the last decade as next-generation technologies of sequencing and multiomics analyses linked the role of the microbial communities to host ... [more ▼]

The human gut microbiota gained tremendous importance in the last decade as next-generation technologies of sequencing and multiomics analyses linked the role of the microbial communities to host physiology and pathophysiology. A growing number of human pathologies and diseases are linked to the gut microbiota. One of the main mechanisms by which the microbiota influences the host is through its interactions with the host immune system. These interactions with both innate and adaptive host intestinal and extraintestinal immunity, although usually commensalistic even mutualistic with the host, in some cases lead to serious health effects. In the case of allogenic hematopoietic stem cell transplantation (allo-HSCT), the disruption of the intestinal microbiota diversity is associated with acute graft-versus-host disease (GvHD). Causing inflammation of the liver, skin, lungs, and the intestine, GvHD occurs in 40-50% of patients undergoing allo-HSCT and results in significant posttransplantation mortality. In this review, we highlight the impact of the gut microbiota on the host immunity in GvHD and the potential of microbiota in alleviation or even prevention of GvHD. [less ▲]

IDENTIFICATION OF A RARE GENE VARIANT THAT IS ASSOCIATED WITH FAMILIAL ALZHEIMER DISEASE AND REGULATES APP EXPRESSION

in Alzheimer's and Dementia: the Journal of the Alzheimer's Association (2017), 13(7, Supplement), 648

Background Genetic mutations leading to familial forms of Alzheimer disease (AD) have so far been reported for a few genes including APP, PSEN1 and PSEN2, UNC5C, PLD3, ABCA7, TTC3, and possibly ADAM10 ... [more ▼]

Background Genetic mutations leading to familial forms of Alzheimer disease (AD) have so far been reported for a few genes including APP, PSEN1 and PSEN2, UNC5C, PLD3, ABCA7, TTC3, and possibly ADAM10. With the advent of whole exome and whole genome sequencing approaches new genes and mutations are likely to be identified. Methods We analyzed the genetic cause of AD in a large multiplex family with an autosomal-dominant pattern of inheritance with LOAD. The family lacked pathogenic mutations of known AD genes. We performed whole-genome sequencing (WGS) in six family members (two affected and four unaffected) and prioritized rare, potential damaging, variants that segregated with disease. Variants were further characterized by subsequent molecular analyzes in human brain and cell culture models. Results We identified a single rare nonsynonymous variant co-segregating with AD. The mutation inhibits pro-protein cleavage and the formation of the active enzyme, thus leading to a loss-of-function of the gene. We further found a strong negative correlation between the identified gene and APP gene expression in human brain and in cells over-expressing the gene. The negative regulation of APP expression was only observed for the wt gene, but not for mutated forms, thus causing beside the loss of enzyme function a decoupling of both APPexpression and subsequent beta-amyloid formation. The identity of the gene will be presented on the conference. Conclusions This novel pathway strongly supports a causative association of the identified gene with the pathogenesis of AD. [less ▲]

Integrated multi-omics of the human gut microbiome in a case study of familial type 1 diabetes.

in Nature Microbiology (2016), 2

The gastrointestinal microbiome is a complex ecosystem with functions that shape human health. Studying the relationship between taxonomic alterations and functional repercussions linked to disease ... [more ▼]

The gastrointestinal microbiome is a complex ecosystem with functions that shape human health. Studying the relationship between taxonomic alterations and functional repercussions linked to disease remains challenging. Here, we present an integrative approach to resolve the taxonomic and functional attributes of gastrointestinal microbiota at the metagenomic, metatranscriptomic and metaproteomic levels. We apply our methods to samples from four families with multiple cases of type 1 diabetes mellitus (T1DM). Analysis of intra- and inter-individual variation demonstrates that family membership has a pronounced effect on the structural and functional composition of the gastrointestinal microbiome. In the context of T1DM, consistent taxonomic differences were absent across families, but certain human exocrine pancreatic proteins were found at lower levels. The associated microbial functional signatures were linked to metabolic traits in distinct taxa. The methodologies and results provide a foundation for future large-scale integrated multi-omic analyses of the gastrointestinal microbiome in the context of host-microbe interactions in human health and disease. [less ▲]