More Like this

1. Mining High-Level Imaging Genetic Associations via Clustering AD Candidate Variants with Similar Brain Association Patterns

   https://doi.org/10.3390/genes13091520  

   Wu, Ruiming; Bao, Jingxuan; Kim, Mansu; Saykin, Andrew J.; Moore, Jason H.; Shen, Li (September 2022, Genes)

   Brain imaging genetics examines associations between imaging quantitative traits (QTs) and genetic factors such as single nucleotide polymorphisms (SNPs) to provide important insights into the pathogenesis of Alzheimer’s disease (AD). The individual level SNP-QT signals are high dimensional and typically have small effect sizes, making them hard to be detected and replicated. To overcome this limitation, this work proposes a new approach that identifies high-level imaging genetic associations through applying multigraph clustering to the SNP-QT association maps. Given an SNP set and a brain QT set, the association between each SNP and each QT is evaluated using a linear regression model. Based on the resulting SNP-QT association map, five SNP–SNP similarity networks (or graphs) are created using five different scoring functions, respectively. Multigraph clustering is applied to these networks to identify SNP clusters with similar association patterns with all the brain QTs. After that, functional annotation is performed for each identified SNP cluster and its corresponding brain association pattern. We applied this pipeline to an AD imaging genetic study, which yielded promising results. For example, in an association study between 54 AD SNPs and 116 amyloid QTs, we identified two SNP clusters with one responsible for amyloid beta clearances and the other regulating amyloid beta formation. These high-level findings have the potential to provide valuable insights into relevant genetic pathways and brain circuits, which can help form new hypotheses for more detailed imaging and genetics studies in independent cohorts. 

   more » « less
2. TA‐RNN: an Attention‐based Time‐aware Recurrent Neural Network Architecture to Predict Progression of Alzheimer’s Disease

   https://doi.org/10.1002/alz.089010  

   Olaimat, Mohammad Al; Bozdag, Serdar; Saeed, Fahad (December 2024, Alzheimer's & Dementia)

   Abstract BackgroundAlzheimer’s Disease (AD) is a widespread neurodegenerative disease with Mild Cognitive Impairment (MCI) acting as an interim phase between normal cognitive state and AD. The irreversible nature of AD and the difficulty in early prediction present significant challenges for patients, caregivers, and the healthcare sector. Deep learning (DL) methods such as Recurrent Neural Networks (RNN) have been utilized to analyze Electronic Health Records (EHR) to model disease progression and predict diagnosis. However, these models do not address some inherent irregularities in EHR data such as irregular time intervals between clinical visits. Furthermore, most DL models are not interpretable. To address these issues, we developed a novel DL architecture called Time‐Aware RNN (TA‐RNN) to predict MCI to AD conversion at the next clinical visit. MethodTA‐RNN comprises of a time embedding layer, attention‐based RNN, and prediction layer based on multi‐layer perceptron (MLP) (Figure 1). For interpretability, a dual‐level attention mechanism within the RNN identifies significant visits and features impacting predictions. TA‐RNN addresses irregular time intervals by incorporating time embedding into longitudinal cognitive and neuroimaging data based on attention weights to create a patient embedding. The MLP, trained on demographic data and the patient embedding predicts AD conversion. TA‐RNN was evaluated on Alzheimer’s Disease Neuroimaging Initiative (ADNI) and National Alzheimer’s Coordinating Center (NACC) datasets based on F2 score and sensitivity. ResultMultiple TA‐RNN models were trained with two, three, five, or six visits to predict the diagnosis at the next visit. In one setup, the models were trained and tested on ADNI. In another setup, the models were trained on the entire ADNI dataset and evaluated on the entire NACC dataset. The results indicated superior performance of TA‐RNN compared to state‐of‐the‐art (SOTA) and baseline approaches for both setups (Figure 2A and 2B). Based on attention weights, we also highlighted significant visits (Figure 3A) and features (Figure 3B) and observed that CDRSB and FAQ features and the most recent visit had highest influence in predictions. ConclusionWe propose TA‐RNN, an interpretable model to predict MCI to AD conversion while handling irregular time intervals. TA‐RNN outperformed SOTA and baseline methods in multiple experiments. 

   more » « less
3. Prioritizing Amyloid Imaging Biomarkers in Alzheimer’s Disease via Learning to Rank

   https://doi.org/10.1007/978-3-030-33226-6_16  

   Peng, B; Ren, Z; Yao, X; Liu, K; Saykin, AJ; Shen, L (October 2019, MBIA 2019: International Workshop on Multimodal Brain Image Analysis)

   We propose an innovative machine learning paradigm enabling precision medicine for AD biomarker discovery. The paradigm tailors the imaging biomarker discovery process to individual characteristics of a given patient. We implement this paradigm using a newly developed learning-to-rank method 𝙿𝙻𝚃𝚁 . The 𝙿𝙻𝚃𝚁 model seamlessly integrates two objectives for joint optimization: pushing up relevant biomarkers and ranking among relevant biomarkers. The empirical study of 𝙿𝙻𝚃𝚁 conducted on the ADNI data yields promising results to identify and prioritize individual-specific amyloid imaging biomarkers based on the individual’s structural MRI data. The resulting top ranked imaging biomarker has the potential to aid personalized diagnosis and disease subtyping. 

   more » « less
4. STAGdb: a 30K SNP genotyping array and Science Gateway for Acropora corals and their dinoflagellate symbionts

   https://doi.org/10.1038/s41598-020-69101-z  

   Kitchen, S. A.; Von Kuster, G.; Kuntz, K. L. Vasquez; Reich, H. G.; Miller, W.; Griffin, S.; Fogarty, Nicole D.; Baums, I. B. (July 2020, Scientific Reports)

   Abstract Standardized identification of genotypes is necessary in animals that reproduce asexually and form large clonal populations such as coral. We developed a high-resolution hybridization-based genotype array coupled with an analysis workflow and database for the most speciose genus of coral,Acropora, and their symbionts. We designed the array to co-analyze host and symbionts based on bi-allelic single nucleotide polymorphisms (SNP) markers identified from genomic data of the two CaribbeanAcroporaspecies as well as their dominant dinoflagellate symbiont,Symbiodinium ‘fitti’.SNPs were selected to resolve multi-locus genotypes of host (called genets) and symbionts (called strains), distinguish host populations and determine ancestry of coral hybrids between Caribbean acroporids. Pacific acroporids can also be genotyped using a subset of the SNP loci and additional markers enable the detection of symbionts belonging to the generaBreviolum, Cladocopium, andDurusdinium. Analytic tools to produce multi-locus genotypes of hosts based on these SNP markers were combined in a workflow called theStandardTools forAcroporidGenotyping (STAG). The STAG workflow and database are contained within a customized Galaxy environment (https://coralsnp.science.psu.edu/galaxy/), which allows for consistent identification of host genet and symbiont strains and serves as a template for the development of arrays for additional coral genera. STAG data can be used to track temporal and spatial changes of sampled genets necessary for restoration planning and can be applied to downstream genomic analyses. Using STAG, we uncover bi-directional hybridization between and population structure within Caribbean acroporids and detect a cryptic Acroporid species in the Pacific. 

   more » « less
5. Neuroimaging and biofluid biomarkers across race and ethnicity in older adults across the spectrum of cognition

   https://doi.org/10.1016/j.arr.2024.102507  

   Wang, Wei-en; Asken, Breton M; DeSimone, Jesse C; Levy, Shellie-Anne; Barker, Warren; Fiala, Jacob A; Velez-Uribe, Idaly; Curiel_Cid, Rosie E; Rósselli, Monica; Marsiske, Michael; et al (November 2024, Ageing Research Reviews)

   Neuroimaging and biofluid biomarkers provide a proxy of pathological changes for Alzheimer’s disease (AD) and are useful in improving diagnosis and assessing disease progression. However, it is not clear how race/ethnicity and different prevalence of AD risks impact biomarker levels. In this narrative review, we survey studies focusing on comparing biomarker differences between non-Hispanic White American(s) (NHW), African American(s) (AA), Hispanic/Latino American(s) (HLA), and Asian American(s) with normal cognition, mild cognitive impairment, and dementia. We found no strong evidence of racial and ethnic differences in imaging biomarkers after controlling for cognitive status and cardiovascular risks. For biofluid biomarkers, in AA, higher levels of plasma Aβ42/Aβ40, and lower levels of CSF total tau and p-tau 181, were observed after controlling for APOE status and comorbidities compared to NHW. Examining the impact of AD risks and comorbidities on biomarkers and their contributions to racial/ethnic differences in cognitive impairment are critical to interpreting biomarkers, understanding their generalizability, and eliminating racial/ethnic health disparities. 

   more » « less