TY - JOUR
T1 - Circular-SWAT for deep learning based diagnostic classification of Alzheimer's disease
T2 - application to metabolome data
AU - Alzheimer's Disease Metabolomics Consortium (ADMC)
AU - Alzheimer's Disease Neuroimaging Initiative (ADNI)
AU - Jo, Taeho
AU - Kim, Junpyo
AU - Bice, Paula
AU - Huynh, Kevin
AU - Wang, Tingting
AU - Arnold, Matthias
AU - Meikle, Peter J.
AU - Giles, Corey
AU - Kaddurah-Daouk, Rima
AU - Saykin, Andrew J.
AU - Nho, Kwangsik
AU - Kueider-Paisley, Alexandra
AU - Doraiswamy, P. Murali
AU - Blach, Colette
AU - Moseley, Arthur
AU - Thompson, Will
AU - St John-Williams, Lisa
AU - Mahmoudiandehkhordi, Siamak
AU - Tenenbaum, Jessica
AU - Welsh-Balmer, Kathleen
AU - Plassman, Brenda
AU - Risacher, Shannon L.
AU - Kastenmüller, Gabi
AU - Han, Xianlin
AU - Baillie, Rebecca
AU - Knight, Rob
AU - Dorrestein, Pieter
AU - Brewer, James
AU - Mayer, Emeran
AU - Labus, Jennifer
AU - Baldi, Pierre
AU - Gupta, Arpana
AU - Fiehn, Oliver
AU - Barupal, Dinesh
AU - Meikle, Peter
AU - Mazmanian, Sarkis
AU - Rader, Dan
AU - Kling, Mitchel
AU - Shaw, Leslie
AU - Trojanowski, John
AU - van Duijin, Cornelia
AU - Nevado-Holgado, Alejo
AU - Bennett, David
AU - Krishnan, Ranga
AU - Keshavarzian, Ali
AU - Vogt, Robin
AU - Ikram, Arfan
AU - Hankemeier, Thomas
AU - Thiele, Ines
AU - Brinton, Roberta
N1 - Publisher Copyright:
© 2023 The Author(s)
PY - 2023/11
Y1 - 2023/11
N2 - Background: Deep learning has shown potential in various scientific domains but faces challenges when applied to complex, high-dimensional multi-omics data. Alzheimer's Disease (AD) is a neurodegenerative disorder that lacks targeted therapeutic options. This study introduces the Circular-Sliding Window Association Test (c-SWAT) to improve the classification accuracy in predicting AD using serum-based metabolomics data, specifically lipidomics. Methods: The c-SWAT methodology builds upon the existing Sliding Window Association Test (SWAT) and utilizes a three-step approach: feature correlation analysis, feature selection, and classification. Data from 997 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI) served as the basis for model training and validation. Feature correlations were analyzed using Weighted Gene Co-expression Network Analysis (WGCNA), and Convolutional Neural Networks (CNN) were employed for feature selection. Random Forest was used for the final classification. Findings: The application of c-SWAT resulted in a classification accuracy of up to 80.8% and an AUC of 0.808 for distinguishing AD from cognitively normal older adults. This marks a 9.4% improvement in accuracy and a 0.169 increase in AUC compared to methods without c-SWAT. These results were statistically significant, with a p-value of 1.04 × 10ˆ-4. The approach also identified key lipids associated with AD, such as Cer(d16:1/22:0) and PI(37:6). Interpretation: Our results indicate that c-SWAT is effective in improving classification accuracy and in identifying potential lipid biomarkers for AD. These identified lipids offer new avenues for understanding AD and warrant further investigation. Funding: The specific funding of this article is provided in the acknowledgements section.
AB - Background: Deep learning has shown potential in various scientific domains but faces challenges when applied to complex, high-dimensional multi-omics data. Alzheimer's Disease (AD) is a neurodegenerative disorder that lacks targeted therapeutic options. This study introduces the Circular-Sliding Window Association Test (c-SWAT) to improve the classification accuracy in predicting AD using serum-based metabolomics data, specifically lipidomics. Methods: The c-SWAT methodology builds upon the existing Sliding Window Association Test (SWAT) and utilizes a three-step approach: feature correlation analysis, feature selection, and classification. Data from 997 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI) served as the basis for model training and validation. Feature correlations were analyzed using Weighted Gene Co-expression Network Analysis (WGCNA), and Convolutional Neural Networks (CNN) were employed for feature selection. Random Forest was used for the final classification. Findings: The application of c-SWAT resulted in a classification accuracy of up to 80.8% and an AUC of 0.808 for distinguishing AD from cognitively normal older adults. This marks a 9.4% improvement in accuracy and a 0.169 increase in AUC compared to methods without c-SWAT. These results were statistically significant, with a p-value of 1.04 × 10ˆ-4. The approach also identified key lipids associated with AD, such as Cer(d16:1/22:0) and PI(37:6). Interpretation: Our results indicate that c-SWAT is effective in improving classification accuracy and in identifying potential lipid biomarkers for AD. These identified lipids offer new avenues for understanding AD and warrant further investigation. Funding: The specific funding of this article is provided in the acknowledgements section.
KW - Alzheimer's disease
KW - Deep learning
KW - Lipidomics
KW - Machine learning
KW - Metabolomics
UR - http://www.scopus.com/inward/record.url?scp=85173462324&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85173462324&partnerID=8YFLogxK
U2 - 10.1016/j.ebiom.2023.104820
DO - 10.1016/j.ebiom.2023.104820
M3 - Article
C2 - 37806288
AN - SCOPUS:85173462324
SN - 2352-3964
VL - 97
JO - EBioMedicine
JF - EBioMedicine
M1 - 104820
ER -