Interpretable deep learning translation of GWAS and multi-omics findings to identify pathobiology and drug repurposing in Alzheimer's disease

Alzheimer's Disease: Combination Therapies and Clinical Trials for Combination Therapy Development

Alzheimer's disease (AD) is a complex multifaceted disease. Recently approved anti-amyloid monoclonal antibodies slow disease progression by approximately 30%, and combination therapy appears necessary to prevent the onset of AD or produce greater slowing of cognitive and functional decline. Combination therapies may address core features, non-specific co-pathology commonly occurring in patients with AD (e.g., inflammation), or non-AD pathologies that may co-occur with AD (e.g., α-synuclein). Combination therapies may be advanced through co-development of more than one new molecular entity or through add-on strategies including an approved agent plus a new molecular entity. Addressing add-on combination therapy is currently urgent since patients on anti-amyloid monoclonal antibodies may be included in clinical trials for experimental agents. Phase 1 information must be generated for each agent in combination drug development. Phase 2 and Phase 3 of add-on therapies may contrast the new molecular entity, the approved agent as standard of care, and the combination. More complex development programs including standard or modified combinatorial designs are required for co-development of two or more new molecular entities. Biomarkers are markedly affected by anti-amyloid monoclonal antibodies, and these effects must be anticipated in add-on trials. Examining target engagement biomarkers and comparing the magnitude and sequence of biomarker changes in those receiving more than one therapy, compared with those on monotherapy, may be informative. Using network-based medicine approaches, computational strategies may identify rational combinations using disease and drug effect network mapping.

Network Medicine: A Potential Approach for Virtual Drug Screening

Traditional drug screening methods typically focus on a single protein target and exhibit limited efficiency due to the multifactorial nature of most diseases, which result from disturbances within complex networks of protein-protein interactions rather than single gene abnormalities. Addressing this limitation requires a comprehensive drug screening strategy. Network medicine is rooted in systems biology and provides a comprehensive framework for understanding disease mechanisms, prevention, and therapeutic innovations. This approach not only explores the associations between various diseases but also quantifies the relationships between disease genes and drug targets within interactome networks, thus facilitating the prediction of drug-disease relationships and enabling the screening of therapeutic drugs for specific complex diseases. An increasing body of research supports the efficiency and utility of network-based strategies in drug screening. This review highlights the transformative potential of network medicine in virtual therapeutic screening for complex diseases, offering novel insights and a robust foundation for future drug discovery endeavors.

GWAS advancements to investigate disease associations and biological mechanisms

Genome-wide association studies (GWAS) have been instrumental in elucidating the genetic architecture of various traits and diseases. Despite the success of GWAS, inherent limitations such as identifying rare and ultra-rare variants, the potential for spurious associations, and in pinpointing causative agents can undermine diagnostic capabilities. This review provides an overview of GWAS and highlights recent advances in genetics that employ a range of methodologies, including Whole Genome Sequencing (WGS), Mendelian Randomization (MR), the Pangenome's high-quality T2T-CHM13 panel, and the Human BioMolecular Atlas Program (HuBMAP), as potential enablers of current and future GWAS research. State of the literature demonstrate the capabilities of these techniques in enhancing the statistical power of GWAS. WGS, with its comprehensive approach, captures the entire genome, surpassing the capabilities of the traditional GWAS technique focused on predefined Single Nucleotide Polymorphism (SNP) sites. The Pangenome's T2T-CHM13 panel, with its holistic approach, aids in the analysis of regions with high sequence identity, such as segmental duplications (SDs). Mendelian Randomization has advanced causative inference, improving clinical diagnostics and facilitating definitive conclusions. Furthermore, spatial biology techniques like HuBMAP, enable 3D molecular mapping of tissues at single-cell resolution, offering insights into pathology of complex traits. This study aims to elucidate and advocate for the increased application of these technologies, highlighting their potential to shape the future of GWAS research.

Tackling neurodegeneration in vitro with omics: a path towards new targets and drugs

Drug discovery is a generally inefficient and capital-intensive process. For neurodegenerative diseases (NDDs), the development of novel therapeutics is particularly urgent considering the long list of late-stage drug candidate failures. Although our knowledge on the pathogenic mechanisms driving neurodegeneration is growing, additional efforts are required to achieve a better and ultimately complete understanding of the pathophysiological underpinnings of NDDs. Beyond the etiology of NDDs being heterogeneous and multifactorial, this process is further complicated by the fact that current experimental models only partially recapitulate the major phenotypes observed in humans. In such a scenario, multi-omic approaches have the potential to accelerate the identification of new or repurposed drugs against a multitude of the underlying mechanisms driving NDDs. One major advantage for the implementation of multi-omic approaches in the drug discovery process is that these overarching tools are able to disentangle disease states and model perturbations through the comprehensive characterization of distinct molecular layers (i.e., genome, transcriptome, proteome) up to a single-cell resolution. Because of recent advances increasing their affordability and scalability, the use of omics technologies to drive drug discovery is nascent, but rapidly expanding in the neuroscience field. Combined with increasingly advanced in vitro models, which particularly benefited from the introduction of human iPSCs, multi-omics are shaping a new paradigm in drug discovery for NDDs, from disease characterization to therapeutics prediction and experimental screening. In this review, we discuss examples, main advantages and open challenges in the use of multi-omic approaches for the in vitro discovery of targets and therapies against NDDs.

Spatial Multi-Omics in Alzheimer's Disease: A Multi-Dimensional Approach to Understanding Pathology and Progression

Alzheimer's Disease (AD) presents a complex neuropathological landscape characterized by hallmark amyloid plaques and neurofibrillary tangles, leading to progressive cognitive decline. Despite extensive research, the molecular intricacies contributing to AD pathogenesis are inadequately understood. While single-cell omics technology holds great promise for application in AD, particularly in deciphering the understanding of different cell types and analyzing rare cell types and transcriptomic expression changes, it is unable to provide spatial distribution information, which is crucial for understanding the pathological processes of AD. In contrast, spatial multi-omics research emerges as a promising and comprehensive approach to analyzing tissue cells, potentially better suited for addressing these issues in AD. This article focuses on the latest advancements in spatial multi-omics technology and compares various techniques. Additionally, we provide an overview of current spatial omics-based research results in AD. These technologies play a crucial role in facilitating new discoveries and advancing translational AD research in the future. Despite challenges such as balancing resolution, increasing throughput, and data analysis, the application of spatial multi-omics holds immense potential in revolutionizing our understanding of human disease processes and identifying new biomarkers and therapeutic targets, thereby potentially contributing to the advancement of AD research.

Artificial intelligence for drug discovery and development in Alzheimer's disease

The complex molecular mechanism and pathophysiology of Alzheimer's disease (AD) limits the development of effective therapeutics or prevention strategies. Artificial Intelligence (AI)-guided drug discovery combined with genetics/multi-omics (genomics, epigenomics, transcriptomics, proteomics, and metabolomics) analysis contributes to the understanding of the pathophysiology and precision medicine of the disease, including AD and AD-related dementia. In this review, we summarize the AI-driven methodologies for AD-agnostic drug discovery and development, including de novo drug design, virtual screening, and prediction of drug-target interactions, all of which have shown potentials. In particular, AI-based drug repurposing emerges as a compelling strategy to identify new indications for existing drugs for AD. We provide several emerging AD targets from human genetics and multi-omics findings and highlight recent AI-based technologies and their applications in drug discovery using AD as a prototypical example. In closing, we discuss future challenges and directions in AI-based drug discovery for AD and other neurodegenerative diseases.

Network medicine informed multi-omics integration identifies drug targets and repurposable medicines for Amyotrophic Lateral Sclerosis

Amyotrophic Lateral Sclerosis (ALS) is a devastating, immensely complex neurodegenerative disease by lack of effective treatments. To date, the challenge to establishing effective treatment for ALS remains formidable, partly due to inadequate translation of existing human genetic findings into actionable ALS-specific pathobiology for subsequent therapeutic development. This study evaluates the feasibility of network medicine methodology via integrating human brain-specific multi-omics data to prioritize drug targets and repurposable treatments for ALS. Using human brain-specific genome-wide quantitative trait loci (x-QTLs) under a network-based deep learning framework, we identified 105 putative ALS-associated genes enriched in various known ALS pathobiological pathways, including regulation of T cell activation, monocyte differentiation, and lymphocyte proliferation. Specifically, we leveraged non-coding ALS loci effects from genome-wide associated studies (GWAS) on brain-specific expression quantitative trait loci (QTL) (eQTL), protein QTLs (pQTL), splicing QTL (sQTL), methylation QTL (meQTL), and histone acetylation QTL (haQTL). Applying network proximity analysis of predicted ALS-associated gene-coding targets and existing drug-target networks under the human protein-protein interactome (PPI) model, we identified a set of potential repurposable drugs (including Diazoxide, Gefitinib, Paliperidone, and Dimethyltryptamine) for ALS. Subsequent validation established preclinical and clinical evidence for top-prioritized repurposable drugs. In summary, we presented a network-based multi-omics framework to identify potential drug targets and repurposable treatments for ALS and other neurodegenerative disease if broadly applied.

Advances in AI and machine learning for predictive medicine

The field of omics, driven by advances in high-throughput sequencing, faces a data explosion. This abundance of data offers unprecedented opportunities for predictive modeling in precision medicine, but also presents formidable challenges in data analysis and interpretation. Traditional machine learning (ML) techniques have been partly successful in generating predictive models for omics analysis but exhibit limitations in handling potential relationships within the data for more accurate prediction. This review explores a revolutionary shift in predictive modeling through the application of deep learning (DL), specifically convolutional neural networks (CNNs). Using transformation methods such as DeepInsight, omics data with independent variables in tabular (table-like, including vector) form can be turned into image-like representations, enabling CNNs to capture latent features effectively. This approach not only enhances predictive power but also leverages transfer learning, reducing computational time, and improving performance. However, integrating CNNs in predictive omics data analysis is not without challenges, including issues related to model interpretability, data heterogeneity, and data size. Addressing these challenges requires a multidisciplinary approach, involving collaborations between ML experts, bioinformatics researchers, biologists, and medical doctors. This review illuminates these complexities and charts a course for future research to unlock the full predictive potential of CNNs in omics data analysis and related fields.

Machine learning models identify predictive features of patient mortality across dementia types

BACKGROUND

Dementia care is challenging due to the divergent trajectories in disease progression and outcomes. Predictive models are needed to flag patients at risk of near-term mortality and identify factors contributing to mortality risk across different dementia types.

METHODS

Here, we developed machine-learning models predicting dementia patient mortality at four different survival thresholds using a dataset of 45,275 unique participants and 163,782 visit records from the U.S. National Alzheimer's Coordinating Center (NACC). We built multi-factorial XGBoost models using a small set of mortality predictors and conducted stratified analyses with dementiatype-specific models.

RESULTS

Our models achieved an area under the receiver operating characteristic curve (AUC-ROC) of over 0.82 utilizing nine parsimonious features for all 1-, 3-, 5-, and 10-year thresholds. The trained models mainly consisted of dementia-related predictors such as specific neuropsychological tests and were minimally affected by other age-related causes of death, e.g., stroke and cardiovascular conditions. Notably, stratified analyses revealed shared and distinct predictors of mortality across eight dementia types. Unsupervised clustering of mortality predictors grouped vascular dementia with depression and Lewy body dementia with frontotemporal lobar dementia.

CONCLUSIONS

This study demonstrates the feasibility of flagging dementia patients at risk of mortality for personalized clinical management. Parsimonious machine-learning models can be used to predict dementia patient mortality with a limited set of clinical features, and dementiatype-specific models can be applied to heterogeneous dementia patient populations.

Artificial intelligence and open science in discovery of disease-modifying medicines for Alzheimer's disease

The high failure rate of clinical trials in Alzheimer's disease (AD) and AD-related dementia (ADRD) is due to a lack of understanding of the pathophysiology of disease, and this deficit may be addressed by applying artificial intelligence (AI) to "big data" to rapidly and effectively expand therapeutic development efforts. Recent accelerations in computing power and availability of big data, including electronic health records and multi-omics profiles, have converged to provide opportunities for scientific discovery and treatment development. Here, we review the potential utility of applying AI approaches to big data for discovery of disease-modifying medicines for AD/ADRD. We illustrate how AI tools can be applied to the AD/ADRD drug development pipeline through collaborative efforts among neurologists, gerontologists, geneticists, pharmacologists, medicinal chemists, and computational scientists. AI and open data science expedite drug discovery and development of disease-modifying therapeutics for AD/ADRD and other neurodegenerative diseases.

The landscape of the methodology in drug repurposing using human genomic data: a systematic review

The process of drug development is expensive and time-consuming. In contrast, drug repurposing can be introduced to clinical practice more quickly and at a reduced cost. Over the last decade, there has been a significant expansion of large biobanks that link genomic data to electronic health record data, public availability of various databases containing biological and clinical information and rapid development of novel methodologies and algorithms in integrating different sources of data. This review aims to provide a thorough summary of different strategies that utilize genomic data to seek drug-repositioning opportunities. We searched MEDLINE and EMBASE databases to identify eligible studies up until 1 May 2023, with a total of 102 studies finally included after two-step parallel screening. We summarized commonly used strategies for drug repurposing, including Mendelian randomization, multi-omic-based and network-based studies and illustrated each strategy with examples, as well as the data sources implemented. By leveraging existing knowledge and infrastructure to expedite the drug discovery process and reduce costs, drug repurposing potentially identifies new therapeutic uses for approved drugs in a more efficient and targeted manner. However, technical challenges when integrating different types of data and biased or incomplete understanding of drug interactions are important hindrances that cannot be disregarded in the pursuit of identifying novel therapeutic applications. This review offers an overview of drug repurposing methodologies, providing valuable insights and guiding future directions for advancing drug repurposing studies.

Deep Trans-Omic Network Fusion for Molecular Mechanism of Alzheimer's Disease

Background

There are various molecular hypotheses regarding Alzheimer's disease (AD) like amyloid deposition, tau propagation, neuroinflammation, and synaptic dysfunction. However, detailed molecular mechanism underlying AD remains elusive. In addition, genetic contribution of these molecular hypothesis is not yet established despite the high heritability of AD.

Objective

The study aims to enable the discovery of functionally connected multi-omic features through novel integration of multi-omic data and prior functional interactions.

Methods

We propose a new deep learning model MoFNet with improved interpretability to investigate the AD molecular mechanism and its upstream genetic contributors. MoFNet integrates multi-omic data with prior functional interactions between SNPs, genes, and proteins, and for the first time models the dynamic information flow from DNA to RNA and proteins.

Results

When evaluated using the ROS/MAP cohort, MoFNet outperformed other competing methods in prediction performance. It identified SNPs, genes, and proteins with significantly more prior functional interactions, resulting in three multi-omic subnetworks. SNP-gene pairs identified by MoFNet were mostly eQTLs specific to frontal cortex tissue where gene/protein data was collected. These molecular subnetworks are enriched in innate immune system, clearance of misfolded proteins, and neurotransmitter release respectively. We validated most findings in an independent dataset. One multi-omic subnetwork consists exclusively of core members of SNARE complex, a key mediator of synaptic vesicle fusion and neurotransmitter transportation.

Conclusions

Our results suggest that MoFNet is effective in improving classification accuracy and in identifying multi-omic markers for AD with improved interpretability. Multi-omic subnetworks identified by MoFNet provided insights of AD molecular mechanism with improved details.

Strategies for translating proteomics discoveries into drug discovery for dementia

Tauopathies, diseases characterized by neuropathological aggregates of tau including Alzheimer's disease and subtypes of frontotemporal dementia, make up the vast majority of dementia cases. Although there have been recent developments in tauopathy biomarkers and disease-modifying treatments, ongoing progress is required to ensure these are effective, economical, and accessible for the globally ageing population. As such, continued identification of new potential drug targets and biomarkers is critical. "Big data" studies, such as proteomics, can generate information on thousands of possible new targets for dementia diagnostics and therapeutics, but currently remain underutilized due to the lack of a clear process by which targets are selected for future drug development. In this review, we discuss current tauopathy biomarkers and therapeutics, and highlight areas in need of improvement, particularly when addressing the needs of frail, comorbid and cognitively impaired populations. We highlight biomarkers which have been developed from proteomic data, and outline possible future directions in this field. We propose new criteria by which potential targets in proteomics studies can be objectively ranked as favorable for drug development, and demonstrate its application to our group's recent tau interactome dataset as an example.

Machine learning in Alzheimer's disease drug discovery and target identification

Alzheimer's disease (AD) stands as a formidable neurodegenerative ailment that poses a substantial threat to the elderly population, with no known curative or disease-slowing drugs in existence. Among the vital and time-consuming stages in the drug discovery process, disease modeling and target identification hold particular significance. Disease modeling allows for a deeper comprehension of disease progression mechanisms and potential therapeutic avenues. On the other hand, target identification serves as the foundational step in drug development, exerting a profound influence on all subsequent phases and ultimately determining the success rate of drug development endeavors. Machine learning (ML) techniques have ushered in transformative breakthroughs in the realm of target discovery. Leveraging the strengths of large dataset analysis, multifaceted data processing, and the exploration of intricate biological mechanisms, ML has become instrumental in the quest for effective AD treatments. In this comprehensive review, we offer an account of how ML methodologies are being deployed in the pursuit of drug discovery for AD. Furthermore, we provide an overview of the utilization of ML in uncovering potential intervention strategies and prospective therapeutic targets for AD. Finally, we discuss the principal challenges and limitations currently faced by these approaches. We also explore the avenues for future research that hold promise in addressing these challenges.

Pan-cancer analysis of promoter activity quantitative trait loci

Altered promoter activity has been generally observed in diverse biological processes, including tumorigenesis. Accumulating evidence suggests that employing a quantitative trait locus mapping approach is effective in comprehending the genetic basis of promoter activity. By utilizing genotype data from The Cancer Genome Atlas and calculating corresponding promoter activity values using proActiv, we systematically evaluated the impact of genetic variants on promoter activity and identified >1.0 million promoter activity quantitative trait loci (paQTLs) as both cis- and trans-acting. Additionally, leveraging data from the genome-wide association study (GWAS) catalog, we discovered >1.3 million paQTLs that overlap with known GWAS linkage disequilibrium regions. Remarkably, ∼9324 paQTLs exhibited significant associations with patient prognosis. Moreover, investigating the impact of promoter activity on >1000 imputed antitumor therapy responses among pan-cancer patients revealed >43 000 million significant associations. Furthermore, ∼25 000 significant associations were identified between promoter activity and immune cell abundance. Finally, a user-friendly data portal, Pancan-paQTL (https://www.hbpding.com/PancanPaQTL/), was constructed for users to browse, search and download data of interest. Pancan-paQTL serves as a comprehensive multidimensional database, enabling functional and clinical investigations into genetic variants associated with promoter activity, drug responses and immune infiltration across multiple cancer types.

Deep learning analysis of UPLC-MS/MS-based metabolomics data to predict Alzheimer's disease

OBJECTIVE

Metabolic biomarkers can potentially inform disease progression in Alzheimer's disease (AD). The purpose of this study is to identify and describe a new set of diagnostic biomarkers for developing deep learning (DL) tools to predict AD using Ultra Performance Liquid Chromatography Mass Spectrometry (UPLC-MS/MS)-based metabolomics data.

METHODS

A total of 177 individuals, including 78 with AD and 99 with cognitive normal (CN), were selected from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort along with 150 metabolomic biomarkers. We performed feature selection using the Least Absolute Shrinkage and Selection Operator (LASSO). The H2O DL function was used to build multilayer feedforward neural networks to predict AD.

RESULTS

The LASSO selected 21 metabolic biomarkers. To develop DL models, the 21 biomarkers identified by LASSO were imported into the H2O package. The data was split into 70% for training and 30% for validation. The best DL model with two layers and 18 neurons achieved an accuracy of 0.881, F1-score of 0.892, and AUC of 0.873. Several metabolomic biomarkers involved in glucose and lipid metabolism, in particular bile acid metabolites, were associated with APOE-ε4 allele and clinical biomarkers (Aβ42, tTau, pTau), cognitive assessments [the Alzheimer's Disease Assessment Scale-cognitive subscale 13 (ADAS13), the Mini-Mental State Examination (MMSE)], and hippocampus volume.

CONCLUSIONS

This study identified a new set of diagnostic metabolomic biomarkers for developing DL tools to predict AD. These biomarkers may help with early diagnosis, prognostic risk stratification, and/or early treatment interventions for patients at risk for AD.

AD-BERT: Using pre-trained language model to predict the progression from mild cognitive impairment to Alzheimer's disease

OBJECTIVE

We develop a deep learning framework based on the pre-trained Bidirectional Encoder Representations from Transformers (BERT) model using unstructured clinical notes from electronic health records (EHRs) to predict the risk of disease progression from Mild Cognitive Impairment (MCI) to Alzheimer's Disease (AD).

METHODS

We identified 3657 patients diagnosed with MCI together with their progress notes from Northwestern Medicine Enterprise Data Warehouse (NMEDW) between 2000 and 2020. The progress notes no later than the first MCI diagnosis were used for the prediction. We first preprocessed the notes by deidentification, cleaning and splitting into sections, and then pre-trained a BERT model for AD (named AD-BERT) based on the publicly available Bio+Clinical BERT on the preprocessed notes. All sections of a patient were embedded into a vector representation by AD-BERT and then combined by global MaxPooling and a fully connected network to compute the probability of MCI-to-AD progression. For validation, we conducted a similar set of experiments on 2563 MCI patients identified at Weill Cornell Medicine (WCM) during the same timeframe.

RESULTS

Compared with the 7 baseline models, the AD-BERT model achieved the best performance on both datasets, with Area Under receiver operating characteristic Curve (AUC) of 0.849 and F1 score of 0.440 on NMEDW dataset, and AUC of 0.883 and F1 score of 0.680 on WCM dataset.

CONCLUSION

The use of EHRs for AD-related research is promising, and AD-BERT shows superior predictive performance in modeling MCI-to-AD progression prediction. Our study demonstrates the utility of pre-trained language models and clinical notes in predicting MCI-to-AD progression, which could have important implications for improving early detection and intervention for AD.

Computational ranking-assisted identification of Plexin-B2 in homotypic and heterotypic clustering of circulating tumor cells in breast cancer metastasis

Metastasis is the cause of over 90% of all deaths associated with breast cancer, yet the strategies to predict cancer spreading based on primary tumor profiles and therefore prevent metastasis are egregiously limited. As rare precursor cells to metastasis, circulating tumor cells (CTCs) in multicellular clusters in the blood are 20-50 times more likely to produce viable metastasis than single CTCs. However, the molecular mechanisms underlying various CTC clusters, such as homotypic tumor cell clusters and heterotypic tumor-immune cell clusters, are yet to be fully elucidated. Combining machine learning-assisted computational ranking with experimental demonstration to assess cell adhesion candidates, we identified a transmembrane protein Plexin- B2 (PB2) as a new therapeutic target that drives the formation of both homotypic and heterotypic CTC clusters. High PB2 expression in human primary tumors predicts an unfavorable distant metastasis-free survival and is enriched in CTC clusters compared to single CTCs in advanced breast cancers. Loss of PB2 reduces formation of homotypic tumor cell clusters as well as heterotypic tumor-myeloid cell clusters in triple-negative breast cancer. Interactions between PB2 and its ligand Sema4C on tumor cells promote homotypic cluster formation, and PB2 binding with Sema4A on myeloid cells (monocytes) drives heterotypic CTC cluster formation, suggesting that metastasizing tumor cells hijack the PB2/Sema family axis to promote lung metastasis in breast cancer. Additionally, using a global proteomic analysis, we identified novel downstream effectors of the PB2 pathway associated with cancer stemness, cell cycling, and tumor cell clustering in breast cancer. Thus, PB2 is a novel therapeutic target for preventing new metastasis.

Network Proximity-based computational pipeline identifies drug candidates for different pathological stages of Alzheimer's disease

Despite the massive investment in Alzheimer's disease (AD), there are still no disease-modifying treatments (DMTs) for AD. One major reason is attributed to the limitation of clinical "one-size-fits-all" approach, since the same AD treatment solely based on clinical diagnosis was unlikely to achieve good clinical efficacy. In recent years, computational approaches based on multiomics data have provided an unprecedented opportunity for drug discovery since they can substantially lower the costs and boost the efficiency. In this study, we intended to identify potential drug candidates for different pathological stages of AD by computationally repurposing Food and Drug Administration (FDA) approved drugs. First, we assembled gene expression data from three different AD pathological stages, which include mild cognitive impairment (MCI) and early and late stages of AD (EAD, LAD). We next quantified the network distances between drug target networks and AD modules by utilizing a network proximity approach, and identified 193 candidates that possessed significant associations with AD. After searching for previous literature evidence, 63 out of 193 (32.6%) predicted drugs were demonstrated to exert therapeutic effects on AD. We further explored the novel mechanism of action (MOA) for these drug candidates by determining the specific brain cells they might function on based on AD patient single cell transcriptomic data. Additionally, we selected several promising candidates that could cross the blood brain barrier together with confirmed neuroprotective effects, and subsequently determined the antioxidative activity of these compounds. Experimental results showed that azathioprine decreased the reactive oxygen species (ROS) and malondialdehyde (MDA) levels and improved the superoxide dismutase (SOD) activity in APP-SH-SY5Y cells. Finally, we deciphered the potential MOA of azathioprine against AD via network analysis and validated several apoptosis-related proteins (Caspase 3, Cleaved Caspase 3, Bax, Bcl2) through western blotting. In summary, this study presented an effective computational strategy utilizing omics data for AD drug repurposing, which provides a new perspective for drug discovery and development.

Identifying Genes Associated with Alzheimer's Disease Using Gene-Based Polygenic Risk Score

BACKGROUND

Except APOE, Alzheimer's disease (AD) associated genes identified in recent large-scale genome-wide association studies (GWAS) had small effects and explained a small portion of heritability. Many AD-associated genes have even smaller effects thereby sub-threshold p-values in large-scale GWAS and remain to be identified. For some AD-associated genes, drug targeting them may have limited efficacies due to their small effect sizes.

OBJECTIVE

The purpose of this study is to identify AD-associated genes with sub-threshold p-values and prioritize drugs targeting AD-associated genes that have large efficacies.

METHODS

We developed a gene-based polygenic risk score (PRS) to identify AD genes. It was calculated using SNPs located within genes and having the same directions of effects in different study cohorts to exclude cohort-specific findings and false positives. Gene co-expression modules and protein-protein interaction networks were used to identify AD-associated genes that interact with multiple other genes, as drugs targeting them have large efficacies via co-regulation or interactions.

RESULTS

Gene-based PRS identified 389 genes with 164 of them not previously reported as AD-associated. These 389 genes explained 56.12% -97.46% SNP heritability; and they were enriched in brain tissues and 164 biological processes, most of which are related to AD and other neurodegenerative diseases. We prioritized 688 drugs targeting 64 genes that were in the same co-expression modules and/or PPI networks.

CONCLUSIONS

Gene-based PRS is a cost-effective way to identify AD-associated genes without substantially increasing the sample size. Co-expression modules and PPI networks can be used to identify drugs having large efficacies.