Abstract
Alzheimer’s Disease (AD) is among the most frequent neuro-degenerative diseases. Early diagnosis is essential for successful disease management and chance to attenuate symptoms by disease modifying drugs. In the past, a number of cerebrospinal fluid (CSF), plasma and neuro-imaging based biomarkers have been proposed. Still, in current clinical practice, AD diagnosis cannot be made until the patient shows clear signs of cognitive decline, which can partially be attributed to the multi-factorial nature of AD. In this work, we integrated genotype information, neuro-imaging as well as clinical data (including neuro-psychological measures) from ~900 normal and mild cognitively impaired (MCI) individuals and developed a highly accurate machine learning model to predict the time until AD is diagnosed. We performed an in-depth investigation of the relevant baseline characteristics that contributed to the AD risk prediction. More specifically, we used Bayesian Networks to uncover the interplay across biological scales between neuro-psychological assessment scores, single genetic variants, pathways and neuro-imaging related features. Together with information extracted from the literature, this allowed us to partially reconstruct biological mechanisms that could play a role in the conversion of normal/MCI into AD pathology. This in turn may open the door to novel therapeutic options in the future.
Highlights
Alzheimer’s Disease (AD) is among the most frequent neuro-degenerative diseases in people above 65 and affects more than 45 Million people worldwide[1]
Together with manually curated cause-effect chains extracted from the literature, this allowed us to partially reconstruct biological mechanisms that could play a role in the conversion of normal/mild cognitively impaired (MCI) into AD pathology
The clinical baseline data of the altogether 924 patients used in this work comprised 73 variables with diagnosis, demographic information, age, gender, education level, neuro-psychological test, MRI and PET scan results, volume measurements of different brain regions as well as 300,000 single nucleotide polymorphisms (SNPs), which are commonly available from both ADNI1 and ADNI2/GO studies
Summary
Alzheimer’s Disease (AD) is among the most frequent neuro-degenerative diseases in people above 65 and affects more than 45 Million people worldwide[1]. A model to predict the time to conversion from 346 MCI into AD based on clinical data, neuro-imaging features and highly restricted genotype information (only 2 SNPs) was developed by Lee et al.[10]. Our model integrated rich genotype information (including newly developed SNP functional pathway impact scores), neuro-imaging (volume measurements of brain regions, PET scan results) as well as clinical data from 900 normal and MCI individuals extracted from the Alzheimer’ s Disease Neuroimaging Initiative (ADNI) (http://adni.loni.usc.edu/), a large scale observational study started in 2004 to evaluate the use of diverse types of biomarkers in clinical practice. A second aim of this work was to better understand the biological mechanisms driving the conversion of normal/MCI into AD pathology, which may open the door to novel therapeutic options To this end, we employed a combination of data driven probabilistic and knowledge driven mechanistic approaches. Together with manually curated cause-effect chains extracted from the literature, this allowed us to partially reconstruct biological mechanisms that could play a role in the conversion of normal/MCI into AD pathology
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
