To identify regulatory non-coding variants associated with Alzheimer’s disease we are performing targeted next-generation sequencing of regulatory genomic elements active in human brain for a cohort of Polish AD patients. We are also looking for disease-associated variants located in the same genomic regions using whole-genome sequencing (WGS) data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the Religious Orders Study and the Rush Memory and Aging Project (ROS/MAP).
Using ADNI and ROS/MAP data we are also exploring the possibilities of applying machine-learning classification algorithms to single-nucleotide polymorphisms (SNPs) from WGS data. We apply the Boruta feature selection algorithm (Kursa and Rudnicki, 2010) and random forest classification algorithm to three datasets: WGS and GWAS data collected by ADNI and WGS data from ROS/MAP. Our results suggest that that the highest possible clssification accuracy is in agreement with estimated AD heritability (around 60%).
People involved:
Marcelina Szczerba (Msc student)
Marlena Osipowicz (Msc student)
Bartek Wilczynski, PhD
Magdalena Machnicka, PhD