A review on omics-based biomarkers discovery for Alzheimer's disease from the bioinformatics perspectives: Statistical approach vs machine learning approach

Tan, Mei Sze; Cheah, Phaik‐Leng; Chin, Ai-Vyrn; Looi, Lai‐Meng; Chang, Siow-Wee

doi:10.1016/j.compbiomed.2021.104947

Cited by 46 publications

(22 citation statements)

References 178 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Phenotypic, lifestyle, and psychosocial characteristics, together with genetic data, and biomarkers would make it possible to create a specific “fingerprint” for each patient, directing towards the creation of personalized interventions and therapies [ 29, 30 ]. In this approach, also the molecular profiles obtained from the multiple omics (genomics, transcriptomics, proteomics, and metabolomics) perspectives are crucial to create a personalized biomarker-targeted treatment [ 31 ].…”

Section: Resultsmentioning

confidence: 99%

“…New and more refined computational analysis techniques, such as machine learning, have been identified to manage the enormous amount of data deriving from the study of omics also in AD. Indeed, machine learning integrates and interprets complex data in scenarios where traditional statistical methods may not be satisfactory [ 31 ]. Recently, Clark et al [ 32 ] performed multi-level CSF omics in a cohort of older adults with normal cognition, MCI, and mild dementia.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Lights and Shadows of Cerebrospinal Fluid Biomarkers in the Current Alzheimer’s Disease Framework

Gallucci

Cenesi

White

et al. 2022

JAD

View full text Add to dashboard Cite

Background: The most significant biomarkers that are included in the Alzheimer’s disease (AD) research framework are amyloid-β plaques deposition, p-tau, t-tau, and neurodegeneration. Although cerebrospinal fluid (CSF) biomarkers are included in the most recent AD research criteria, their use is increasing in the routine clinical practice and is applied also to the preclinical phases of AD, including mild cognitive impairment. The role of these biomarkers is still unclear concerning the preclinical stage of AD diagnosis, the CSF methodology, and the costs-benefits of the biomarkers’ tests. The controversies regarding the use of biomarkers in the clinical practice are related to the concepts of analytical validity, clinical validity, and clinical utility and to the question of whether they are able to diagnose AD without the support of AD clinical phenotypes. Objective: The objective of the present work is to expose the strengths and weaknesses of the use of CSF biomarkers in the diagnosis of AD in a clinical context. Methods: We used PubMed as main source for articles published and the final reference list was generated on the basis of relevance to the topics covered in this work. Results: The use of CSF biomarkers for AD diagnosis is certainly important but its indication in routine clinical practice, especially for prodromal conditions, needs to be regulated and also contextualized considering the variety of possible clinical AD phenotypes. Conclusion: We suggest that the diagnosis of AD should be understood both as clinical and pathological.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Lights and Shadows of Cerebrospinal Fluid Biomarkers in the Current Alzheimer’s Disease Framework

Gallucci

Cenesi

White

et al. 2022

JAD

View full text Add to dashboard Cite

show abstract

“…Several lately published studies provide tools for an efficient early diagnosis or even prognosis of AD development or progression based on the Bayesian Inference or Machine Learning Techniques ( Mantzavinos and Alexiou, 2017 ; Ali et al, 2019 ; Ashraf and Alexiou, 2019 , 2022 ; De Velasco Oriol et al, 2019 ; Alexiou et al, 2020 ; Tan et al, 2021 ; Mirzaei and Adeli, 2022 ).…”

Section: Discussionmentioning

confidence: 99%

Dietary Alterations in Impaired Mitochondrial Dynamics Due to Neurodegeneration

Ashraf

Chatzichronis²,

Αλεξίου³

et al. 2022

Front. Aging Neurosci.

View full text Add to dashboard Cite

Alzheimer’s disease is still an incurable disease with significant social and economic impact globally. Nevertheless, newly FDA-approved drugs and non-pharmacological techniques may offer efficient disease treatments. Furthermore, it is widely accepted that early diagnosis or even prognosis of Alzheimer’s disease using advanced computational tools could offer a compelling alternative way of management. In addition, several studies have presented an insight into the role of mitochondrial dynamics in Alzheimer’s development. In combination with diverse dietary and obesity-related diseases, mitochondrial bioenergetics may be linked to neurodegeneration. Considering the probabilistic expectations of Alzheimer’s disease development or progression due to specific risk factors or biomarkers, we designed a Bayesian model to formulate the impact of diet-induced obesity with an impaired mitochondrial function and altered behavior. The applied probabilities are based on clinical trials globally and are continuously subject to updating and redefinition. The proposed multiparametric model combines various data types based on uniform probabilities. The program simulates all the variables with a uniform distribution in a sample of 1000 patients. First, the program initializes the variable age (30–95) and the four different diet types (“HFO_diet,” “Starvation,” “HL_diet,” “CR”) along with the factors that are related to prodromal or mixed AD (ATP, MFN1, MFN2, DRP1, FIS1, Diabetes, Oxidative_Stress, Hypertension, Obesity, Depression, and Physical_activity). Besides the known proteins related to mitochondrial dynamics, our model includes risk factors like Age, Hypertension, Oxidative Stress, Obesity, Depression, and Physical Activity, which are associated with Prodromal Alzheimer’s. The outcome is the disease progression probability corresponding to a random individual ID related to diet choices and mitochondrial dynamics parameters. The proposed model and the programming code are adjustable to different parameters and values. The program is coded and executed in Python and is fully and freely available for research purposes and testing the correlation between diet type and Alzheimer’s disease progression regarding various risk factors and biomarkers.

show abstract

“…Other approaches, which are vastly more computationally complex, include reducing the number of dimensions (e.g., through principal components, latent classes, k -means clustering, etc.) and variable selection steps (e.g., LASSO or stepwise regression) [ 38 , 47 , 48 , 50 ]. These approaches often sub-set the full dataset into training and validation subsets and then use cross-validation to select tuning parameters from the training subset, validate the parameters on the validation subset, and then apply the parameters to the entire dataset [ 38 , 47 , 48 , 50 ].…”

Section: Introductionmentioning

confidence: 99%

“…and variable selection steps (e.g., LASSO or stepwise regression) [ 38 , 47 , 48 , 50 ]. These approaches often sub-set the full dataset into training and validation subsets and then use cross-validation to select tuning parameters from the training subset, validate the parameters on the validation subset, and then apply the parameters to the entire dataset [ 38 , 47 , 48 , 50 ].…”

Section: Introductionmentioning

confidence: 99%

Machine Learning Approaches for Measuring Neighborhood Environments in Epidemiologic Studies

2022

View full text Add to dashboard Cite

Purpose of review Innovations in information technology, initiatives by local governments to share administrative data, and growing inventories of data available from commercial data aggregators have immensely expanded the information available to describe neighborhood environments, supporting an approach to research we call Urban Health Informatics. This review evaluates the application of machine learning to this new wealth of data for studies of the effects of neighborhood environments on health. Recent findings Prominent machine learning applications in this field include automated image analysis of archived imagery such as Google Street View images, variable selection methods to identify neighborhood environment factors that predict health outcomes from large pools of exposure variables, and spatial interpolation methods to estimate neighborhood conditions across large geographic areas. Summary In each domain, we highlight successes and cautions in the application of machine learning, particularly highlighting legal issues in applying machine learning approaches to Google’s geo-spatial data.

show abstract

A review on omics-based biomarkers discovery for Alzheimer's disease from the bioinformatics perspectives: Statistical approach vs machine learning approach

Cited by 46 publications

References 178 publications

Lights and Shadows of Cerebrospinal Fluid Biomarkers in the Current Alzheimer’s Disease Framework

Lights and Shadows of Cerebrospinal Fluid Biomarkers in the Current Alzheimer’s Disease Framework

Dietary Alterations in Impaired Mitochondrial Dynamics Due to Neurodegeneration

Machine Learning Approaches for Measuring Neighborhood Environments in Epidemiologic Studies

Contact Info

Product

Resources

About