A robust and interpretable machine learning approach using multimodal biological data to predict future pathological tau accumulation.
Ontology highlight
ABSTRACT:
SUBMITTER: Giorgio J
PROVIDER: S-EPMC8989879 | biostudies-literature | 2022 Apr
REPOSITORIES: biostudies-literature
Publications
A robust and interpretable machine learning approach using multimodal biological data to predict future pathological tau accumulation.
Giorgio Joseph J Jagust William J WJ Baker Suzanne S Landau Susan M SM Tino Peter P Kourtzi Zoe Z
Nature communications 20220407 1
The early stages of Alzheimer's disease (AD) involve interactions between multiple pathophysiological processes. Although these processes are well studied, we still lack robust tools to predict individualised trajectories of disease progression. Here, we employ a robust and interpretable machine learning approach to combine multimodal biological data and predict future pathological tau accumulation. In particular, we use machine learning to quantify interactions between key pathological markers ...[more]
Similar Datasets
Project description:BackgroundDistal sensorimotor polyneuropathy (DSPN) is a common neurological disorder in elderly adults and people with obesity, prediabetes and diabetes and is associated with high morbidity and premature mortality. DSPN is a multifactorial disease and not fully understood yet.MethodsHere, we developed the Interpretable Multimodal Machine Learning (IMML) framework for predicting DSPN prevalence and incidence based on sparse multimodal data. Exploiting IMMLs interpretability further empowered biomarker identification. We leveraged the population-based KORA F4/FF4 cohort including 1091 participants and their deep multimodal characterisation, i.e. clinical data, genomics, methylomics, transcriptomics, proteomics, inflammatory proteins and metabolomics.ResultsClinical data alone is sufficient to stratify individuals with and without DSPN (AUROC = 0.752), whilst predicting DSPN incidence 6.5 ± 0.2 years later strongly benefits from clinical data complemented with two or more molecular modalities (improved ΔAUROC > 0.1, achieved AUROC of 0.714). Important and interpretable features of incident DSPN prediction include up-regulation of proinflammatory cytokines, down-regulation of SUMOylation pathway and essential fatty acids, thus yielding novel insights in the disease pathophysiology.ConclusionsThese may become biomarkers for incident DSPN, guide prevention strategies and serve as proof of concept for the utility of IMML in studying complex diseases.
| S-EPMC11649904 | biostudies-literature
Project description:Clinical trials with anti-tau drugs will need to target individuals at risk of accumulating tau. Our objective was to identify variables available in a research setting that predict future rates of tau PET accumulation separately among individuals who were either cognitively unimpaired or cognitively impaired. All 337 participants had: a baseline study visit with MRI, amyloid PET, and tau PET exams, at least one follow-up tau PET exam; and met clinical criteria for membership in one of two clinical diagnostic groups: cognitively unimpaired (n = 203); or cognitively impaired (n = 134, a combined group of participants with either mild cognitive impairment or dementia with Alzheimer's clinical syndrome). Our primary analyses were in these two clinical groups; however, we also evaluated subgroups dividing the unimpaired group by normal/abnormal amyloid PET and the impaired group by clinical phenotype (mild cognitive impairment, amnestic dementia, and non-amnestic dementia). Linear mixed effects models were used to estimate associations between age, sex, education, APOE genotype, amyloid and tau PET standardized uptake value ratio (SUVR), cognitive performance, cortical thickness, and white matter hyperintensity volume at baseline, and the rate of subsequent tau PET accumulation. Log-transformed tau PET SUVR was used as the response and rates were summarized as annual per cent change. A temporal lobe tau PET meta-region of interest was used. In the cognitively unimpaired group, only higher baseline amyloid PET was a significant independent predictor of higher tau accumulation rates (P < 0.001). Higher rates of tau accumulation were associated with faster rates of cognitive decline in the cognitively unimpaired subgroup with abnormal amyloid PET (P = 0.03), but among the subgroup with normal amyloid PET. In the cognitively impaired group, younger age (P = 0.02), higher baseline amyloid PET (P = 0.05), APOE ε4 (P = 0.05), and better cognitive performance (P = 0.05) were significant independent predictors of higher tau accumulation rates. Among impaired individuals, faster cognitive decline was associated with faster rates of tau accumulation (P = 0.01). While we examined many possible predictor variables, our results indicate that screening of unimpaired individuals for potential inclusion in anti-tau trials may be straightforward because the only independent predictor of high tau rates was amyloidosis. In cognitively impaired individuals, imaging and clinical variables consistent with early onset Alzheimer's disease phenotype were associated with higher rates of tau PET accumulation suggesting this may be a highly advantageous group in which to conduct proof-of-concept clinical trials that target tau-related mechanisms. The nature of the dementia phenotype (amnestic versus non-amnestic) did not affect this conclusion.
| S-EPMC7586089 | biostudies-literature
Project description:Tau aggregation is central to the pathogenesis of a large group of neurodegenerative diseases termed tauopathies, but it is still unclear in which way neurons respond to tau pathology and how tau accumulation leads to neurodegeneration. A striking neuron-specific response to tau pathology is presented by granulovacuolar degeneration bodies (GVBs), lysosomal structures that accumulate specific cargo in a dense core. Here we employed different tau aggregation models in primary neurons to investigate which properties of pathological tau assemblies affect GVB accumulation using a combination of confocal microscopy, transmission electron microscopy, and quantitative automated high-content microscopy. Employing GFP-tagged and untagged tau variants that spontaneously form intraneuronal aggregates, we induced pathological tau assemblies with a distinct subcellular localization, morphology, and ultrastructure depending on the presence or absence of the GFP tag. The quantification of the GVB load in the different models showed that an increased GVB accumulation is associated with the untagged tau aggregation model, characterized by shorter and more randomly distributed tau filaments in the neuronal soma. Our data indicate that tau aggregate structure and/or subcellular localization may be key determinants of GVB accumulation.
| S-EPMC10341458 | biostudies-literature
Project description:Machine learning tools have proven useful across biological disciplines, allowing researchers to draw conclusions from large datasets, and opening up new opportunities for interpreting complex and heterogeneous biological data. Alongside the rapid growth of machine learning, there have also been growing pains: some models that appear to perform well have later been revealed to rely on features of the data that are artifactual or biased; this feeds into the general criticism that machine learning models are designed to optimize model performance over the creation of new biological insights. A natural question arises: how do we develop machine learning models that are inherently interpretable or explainable? In this manuscript, we describe the SWIF(r) reliability score (SRS), a method building on the SWIF(r) generative framework that reflects the trustworthiness of the classification of a specific instance. The concept of the reliability score has the potential to generalize to other machine learning methods. We demonstrate the utility of the SRS when faced with common challenges in machine learning including: 1) an unknown class present in testing data that was not present in training data, 2) systemic mismatch between training and testing data, and 3) instances of testing data that have missing values for some attributes. We explore these applications of the SRS using a range of biological datasets, from agricultural data on seed morphology, to 22 quantitative traits in the UK Biobank, and population genetic simulations and 1000 Genomes Project data. With each of these examples, we demonstrate how the SRS can allow researchers to interrogate their data and training approach thoroughly, and to pair their domain-specific knowledge with powerful machine-learning frameworks. We also compare the SRS to related tools for outlier and novelty detection, and find that it has comparable performance, with the advantage of being able to operate when some data are missing. The SRS, and the broader discussion of interpretable scientific machine learning, will aid researchers in the biological machine learning space as they seek to harness the power of machine learning without sacrificing rigor and biological insight.
| S-EPMC10249903 | biostudies-literature
Project description:IntroductionAlzheimer's disease is partially characterized by the progressive accumulation of aggregated tau-containing neurofibrillary tangles. Although the association between accumulated tau, neurodegeneration, and cognitive decline is critical for disease understanding and clinical trial design, we still lack robust tools to predict individualized trajectories of tau accumulation. Our objective was to assess whether brain imaging biomarkers of flortaucipir-positron emission tomography (PET), in combination with clinical and genomic measures, could predict future pathological tau accumulation.MethodsWe quantified the disease profile of participants (N = 276) using a comprehensive set of descriptors, including clinical/demographic (age, diagnosis, amyloid status, sex, race, ethnicity), genetic (apolipoprotein E [APOE]-ε4), and flortaucipir-PET imaging measures (regional flortaucipir standardized uptake value ratio [SUVr] and comprehensive radiomic texture features extracted from Automated Anatomical Labeling template regions). We trained an AdaBoost machine learning algorithm in a 2:1 split train-test configuration to derive a prognostic index that (i) stratifies individualized brain regions including global (AD-signature region) and lobar regions (frontal, occipital, parietal, temporal) into stable/slow- and fast-progressors based on future tau accumulation, and (ii) forecasts individualized regional annualized-rate-of-change in flortaucipir-PET SUVr. Further, we developed an adaptive model incorporating flortaucipir-PET measurements from the baseline and intermediate timepoints to predict annualized-rate-of-change.ResultsIn binary classification for predicting stable/slow- versus fast-progressors, the area-under-the-receiver-operating-characteristic curve was 0.86 in the AD-signature region and 0.83, 0.82, 0.84, and 0.83 in frontal, occipital, parietal, and temporal regions, respectively. The trained models successfully predicted annualized-rate-of-change of flortaucipir-PET regional flortaucipir SUVr in AD-signature and lobar regions (Pearson-correlation [R]: AD-signature = 0.73; frontal = 0.73; occipital = 0.71; parietal = 0.70; temporal = 0.69). The models' performance in predicting annualized-rate-of-change slightly increased when imaging features from intermediate timepoints were used in the adaptive setting (R: AD-signature = 0.79; frontal = 0.87; occipital = 0.83; parietal = 0.74; temporal = 0.82).DiscussionTaken together, our results propose a robust approach to predict future tau accumulation that may improve the ability to enroll, stratify, and gauge efficacy in clinical trial participants.HighlightsMachine learning predicts the future rate of tau accumulation in Alzheimer's disease.Tau prediction in lobar/global regions benefits from diverse multimodal features.This prognostic index can serve as a sensitive tool for patient stratification.
| S-EPMC11694527 | biostudies-literature
Project description:Pathological aggregates of phosphorylated TDP-43 characterize amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD-TDP), two devastating groups of neurodegenerative disease. Kinase hyperactivity may be a consistent feature of ALS and FTLD-TDP, as phosphorylated TDP-43 is not observed in the absence of neurodegeneration. By examining changes in TDP-43 phosphorylation state, we have identified kinases controlling TDP-43 phosphorylation in a C. elegans model of ALS. In this kinome-wide survey, we identified homologs of the tau tubulin kinases 1 and 2 (TTBK1 and TTBK2), which were also identified in a prior screen for kinase modifiers of TDP-43 behavioral phenotypes. Using refined methodology, we demonstrate TTBK1 and TTBK2 directly phosphorylate TDP-43 in vitro and promote TDP-43 phosphorylation in mammalian cultured cells. TTBK1/2 overexpression drives phosphorylation and relocalization of TDP-43 from the nucleus to cytoplasmic inclusions reminiscent of neuropathologic changes in disease states. Furthermore, protein levels of TTBK1 and TTBK2 are increased in frontal cortex of FTLD-TDP patients, and TTBK1 and TTBK2 co-localize with TDP-43 inclusions in ALS spinal cord. These kinases may represent attractive targets for therapeutic intervention for TDP-43 proteinopathies such as ALS and FTLD-TDP.
| S-EPMC4256087 | biostudies-literature
Project description:ObjectivesTo develop a computed tomography (CT) radiomics-based interpretable machine learning (ML) model to predict the pathological grade of pancreatic neuroendocrine tumors (pNETs) in a non-invasive manner.MethodsPatients with pNETs who underwent contrast-enhanced abdominal CT between 2010 and 2022 were included in this retrospective study. Radiomics features were extracted, and five radiomics-based ML models, namely logistic regression (LR), random forest (RF), support vector machine (SVM), XGBoost, and GaussianNB, were developed. The performance of these models was evaluated using a time-independent testing set, and metrics such as sensitivity, specificity, accuracy, and the area under the receiver operating characteristic curve (AUC) were calculated. The accuracy of the radiomics model was compared to that of needle biopsy. The Shapley Additive Explanation (SHAP) tool and the correlation between radiomics and biological features were employed to explore the interpretability of the model.ResultsA total of 122 patients (mean age: 50 ± 14 years; 53 male) were included in the training set, whereas 100 patients (mean age: 48 ± 13 years; 50 male) were included in the testing set. The AUCs for LR, SVM, RF, XGBoost, and GaussianNB were 0.758, 0.742, 0.779, 0.744, and 0.745, respectively, with corresponding accuracies of 73.0%, 70.0%, 77.0%, 71.9%, and 72.9%. The SHAP tool identified two features of the venous phase as the most significant, which showed significant differences among the Ki-67 index or mitotic count subgroups (p < 0.001).ConclusionsAn interpretable radiomics-based RF model can effectively differentiate between G1 and G2/3 of pNETs, demonstrating favorable interpretability.Clinical relevance statementThe radiomics-based interpretable model developed in this study has significant clinical relevance as it offers a non-invasive method for assessing the pathological grade of pancreatic neuroendocrine tumors and holds promise as an important complementary tool to traditional tissue biopsy.Key points• A radiomics-based interpretable model was developed to predict the pathological grade of pNETs and compared with preoperative needle biopsy in terms of accuracy. • The model, based on CT radiomics, demonstrated favorable interpretability. • The radiomics model holds potential as a valuable complementary technique to preoperative needle biopsy; however, it should not be considered a replacement for biopsy.
| S-EPMC10873440 | biostudies-literature
Project description:A major hurdle to clinical translation of brain-machine interfaces (BMIs) is that current decoders, which are trained from a small quantity of recent data, become ineffective when neural recording conditions subsequently change. We tested whether a decoder could be made more robust to future neural variability by training it to handle a variety of recording conditions sampled from months of previously collected data as well as synthetic training data perturbations. We developed a new multiplicative recurrent neural network BMI decoder that successfully learned a large variety of neural-to-kinematic mappings and became more robust with larger training data sets. Here we demonstrate that when tested with a non-human primate preclinical BMI model, this decoder is robust under conditions that disabled a state-of-the-art Kalman filter-based decoder. These results validate a new BMI strategy in which accumulated data history are effectively harnessed, and may facilitate reliable BMI use by reducing decoder retraining downtime.
| S-EPMC5159828 | biostudies-literature
Project description:BackgroundTranslation of predictive and prognostic image-based learning models to clinical applications is challenging due in part to their lack of interpretability. Some deep-learning-based methods provide information about the regions driving the model output. Yet, due to the high-level abstraction of deep features, these methods do not completely solve the interpretation challenge. In addition, low sample size cohorts can lead to instabilities and suboptimal convergence for models involving a large number of parameters such as convolutional neural networks.PurposeHere, we propose a method for designing radiomic models that combines the interpretability of handcrafted radiomics with a sub-regional analysis.Materials and methodsOur approach relies on voxel-wise engineered radiomic features with average global aggregation and logistic regression. The method is illustrated using a small dataset of 51 soft tissue sarcoma (STS) patients where the task is to predict the risk of lung metastasis occurrence during the follow-up period.ResultsUsing positron emission tomography/computed tomography and two magnetic resonance imaging sequences separately to build two radiomic models, we show that our approach produces quantitative maps that highlight the signal that contributes to the decision within the tumor region of interest. In our STS example, the analysis of these maps identified two biological patterns that are consistent with STS grading systems and knowledge: necrosis development and glucose metabolism of the tumor.ConclusionsWe demonstrate how that method makes it possible to spatially and quantitatively interpret radiomic models amenable to sub-regions identification and biological interpretation for patient stratification.
| S-EPMC9325536 | biostudies-literature
Project description:Dysregulated proteostasis is a key feature of a variety of neurodegenerative disorders. In Alzheimer's disease (AD), progression of symptoms closely correlates with spatiotemporal progression of Tau aggregation, with "early" oligomeric Tau forms rather than mature neurofibrillary tangles (NFTs) considered to be pathogenetic culprits. The ubiquitin-proteasome system (UPS) controls degradation of soluble normal and abnormally folded cytosolic proteins. The UPS is affected in AD and is identified by genomewide association study (GWAS) as a risk pathway for AD. The UPS is determined by balanced regulation of ubiquitination and deubiquitination. In this work, we performed isobaric tags for relative and absolute quantitation (iTRAQ)-based Tau interactome mapping to gain unbiased insight into Tau pathophysiology and to identify novel Tau-directed therapeutic targets. Focusing on Tau deubiquitination, we here identify Otub1 as a Tau-deubiquitinating enzyme. Otub1 directly affected Lys48-linked Tau deubiquitination, impairing Tau degradation, dependent on its catalytically active cysteine, but independent of its noncanonical pathway modulated by its N-terminal domain in primary neurons. Otub1 strongly increased AT8-positive Tau and oligomeric Tau forms and increased Tau-seeded Tau aggregation in primary neurons. Finally, we demonstrated that expression of Otub1 but not its catalytically inactive form induced pathological Tau forms after 2 months in Tau transgenic mice in vivo, including AT8-positive Tau and oligomeric Tau forms. Taken together, we here identified Otub1 as a Tau deubiquitinase in vitro and in vivo, involved in formation of pathological Tau forms, including small soluble oligomeric forms. Otub1 and particularly Otub1 inhibitors, currently under development for cancer therapies, may therefore yield interesting novel therapeutic avenues for Tauopathies and AD.
| S-EPMC5390007 | biostudies-literature