Back

Identification of novel variants, genes and pathways potentially linked to Parkinson's disease using machine learning

Yu, E.; Lariviere, R.; Thomas, R. A.; Liu, L.; Senkevich, K.; Rahayel, S.; Trempe, J.-F.; Fon, E. A.; Gan-Or, Z.

2023-06-27 genetic and genomic medicine
10.1101/2023.06.20.23291658 medRxiv
Show abstract

There are 78 loci associated with Parkinsons disease (PD) in the most recent genome-wide association study (GWAS), yet the specific genes driving these associations are mostly unknown. Herein, we aimed to nominate the top candidate gene from each PD locus, and identify variants and pathways potentially involved in PD. We trained a machine learning model to predict PD-associated genes from GWAS loci using genomic, transcriptomic, and epigenomic data from brain tissues and dopaminergic neurons. We nominated candidate genes in each locus, identified novel pathways potentially involved in PD, such as the inositol phosphate biosynthetic pathway (INPP5F, IP6K2, ITPKB, PPIP5K2). Specific common coding variants in SPNS1 and MLX may be involved in PD, and burden tests of rare variants further support that CNIP3, LSM7, NUCKS1 and the polyol/inositol phosphate biosynthetic pathway are associated with PD. Functional studies are needed to further analyze the involvements of these genes and pathways in PD.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.