Prediction of Alzheimer’s disease using blood gene expression data

Abstract

Identification of AD (Alzheimer's disease)-related genes obtained from blood samples is crucial for early AD diagnosis. We used three public datasets, ADNI, AddNeuroMed1 (ANM1), and ANM2, for this study. Five feature selection methods and five classifiers were used to curate AD-related genes and discriminate AD patients, respectively. In the internal validation (five-fold cross-validation within each dataset), the best average values of the area under the curve (AUC) were 0.657, 0.874, and 0.804 for ADNI, ANMI, and ANM2, respectively. In the external validation (training and test sets from different datasets), the best AUCs were 0.697 (training: ADNI to testing: ANM1), 0.764 (ADNI to ANM2), 0.619 (ANM1 to ADNI), 0.79 (ANM1 to ANM2), 0.655 (ANM2 to ADNI), and 0.859 (ANM2 to ANM1), respectively. These results suggest that although the classification performance of ADNI is relatively lower than that of ANM1 and ANM2, classifiers trained using blood gene expression can be used to classify AD for other data sets. In addition, pathway analysis showed that AD-related genes were enriched with inflammation, mitochondria, and Wnt signaling pathways. Our study suggests that blood gene expression data are useful in predicting the AD classification. Alzheimer's disease (AD), the most common form of dementia, is estimated to affect in 13.8 million individuals in the United States (US), with 7.0 million being aged 85 years or older by 2050 1. Based on the National Institute of Neurological, Communicative Disorders, and Stroke and Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) criteria in 1985, probable or possible AD was diagnosed based on subjective symptoms and questionnaires 2. Recently, the transition from symptom-based to pathophysiology-based AD diagnosis showed that AD diagnosis is mainly based on structural brain changes (MRI), molecular neuroimaging changes (positron emission tomography imaging), and alterations in cerebral spinal fluid biomarkers 3. Although the elucidation of the biological basis of AD has resulted in many advancements 3 , early diagnostic detection of AD remains challenging. Recent advances in biotechnology have led to full-scale analyses of the genome, transcriptome, and epigenome rather than focusing on a few biomarkers. A large-scale genome-wide association study (GWAS) of 2,032 individuals with AD and 5,328 controls was presented in 2009 and it identified variants at CLU and CR1, which were associated with AD 4. Additionally, a meta-analysis of four previously reported GWAS datasets (17,008 AD cases, 37,154 controls) yielded 11 new loci of susceptibility to AD 5. Recently, Xu et al. constructed an AlzData database integrating data from GWAS, eQTL, interactome, and laboratory experiments 6 , which provides all human genes with scores for association with AD, called the Convergent Functional Genomics (CFG) score 7,8. In recent years, two large multi-center studies were conducted to identify biomarkers for early AD diagnosis and MCI progression to AD: the Europe-based ANM and ...

Predicting Clinical Dementia Rating Using Blood RNA Levels

“…MAPK14 regulates immunological responses and integral in the production of chemokines and cytokines in astrocytes [ 27 ]. Both genes are involved in immune response, and support previous research indicating association of AD with differential expression in gene integral to the immune system [ 16 ]. Additionally, MAPK14 is located 4 Mbp downstream from CLIC1 on chromosome 6, and the proximity to CLIC1 may cause a false positive significant p -value due to gene interactions or linkage disequilibrium.…”

“…Previous studies have used blood mRNA levels to predict cognitive decline with mixed results. Lee and Lee [ 16 ] identified significant differences in expression in genes that were enriched in inflammatory and immune pathways, although the predictive power varied significantly between studies. Similarly, differential expression in genes implicated in other autoimmune diseases have previously been linked to AD [ 17 ].…”

Accurate Blood-Based Diagnostic Biosignatures for Alzheimer’s Disease via Automated Machine Learning

“…Lee and Lee produced three mRNA-biosignatures via logistic regression, L1-regularized logistic regression, random forest, support vector machine, and deep neural network using three independent datasets reaching AUC 0.657, 0.874, and 0.804, respectively. [ 27 ] Li et al, using the same 38,327-feature mRNA transcriptomic dataset of our study (containing all the available samples) and trying “manually” different ML approaches, including support vector machines, random forest and logistic Ridge Regression models, produced a three-mRNA biosignature which discriminated between AD patients and controls with 0.860 AUC [ 28 ]. AutoML analysis of this dataset here led to six statistically equivalent biosignatures via a Classification Random Forests model of similar performance.…”

Functional Genetic Biomarkers of Alzheimer’s Disease and Gene Expression from Peripheral Blood

“…These studies utilize gene expression values from the brain tissue in AD subjects. Overall, this means that the analysis was largely based on samples from biopsies or autopsies [54]. This is undesirable because these lab-based samples and analyses are difficult to extrapolate to a clinical setting.…”