Back

NeuroSCORE: A Genome-wide Omics-Based Model to Identify Candidate Disease Genes of the Central Nervous System

Davis, K. W.; Bilancia, C. G.; Martin, M.; Vanzo, R.; Rimmasch, M.; Hom, Y.; Uddin, M.; Serrano, M.

2021-02-06 genomics
10.1101/2021.02.04.429640 bioRxiv
Show abstract

To identify and prioritize candidate disease genes of the central nervous system (CNS) we created the Neurogenetic Systematic Correlation of Omics-Related Evidence (NeuroSCORE). We used five genome-wide metrics highly associated with neurological phenotypes to score 19,598 protein-coding genes. Genes scored one point per metric, resulting in a range of scores from 0-5. Approximately 13,000 genes were then paired with phenotype data from the Online Mendelian Inheritance in Man (OMIM) database. We used logistic regression to determine the odds ratio of each metric and compared genes scoring 1+ to cause a known CNS-related phenotype compared to genes that scored zero. We tested NeuroSCORE using microarray copy number variants (CNVs) in case-control cohorts, mouse model phenotype data, and gene ontology (GO) and pathway analyses. NeuroSCORE identified 8,296 genes scored [&ge;]1, of which 1,580 are "high scoring" genes (scores [&ge;]3). High scoring genes are significantly associated with CNS phenotypes (OR=5.5, p<2x10-16), enriched in case CNVs, and enriched in mouse ortholog genes associated with behavioral and nervous system abnormalities. GO and pathway analyses showed high scoring genes were enriched in chromatin remodeling, mRNA splicing, dendrite development, and neuron projection. OMIM has no phenotype for 1,062 high scoring genes (67%). Top scoring genes include ANKRD17, CCAR1, CLASP1, DOCK9, EIF4G2, G3BP2, GRIA1, MAP4K4, MARK2, PCBP2, RNF145, SF1, SYNCRIP, TNPO2, and ZSWIM8. NeuroSCORE identifies and prioritizes CNS-disease candidate genes, many not yet associated with any phenotype in OMIM. These findings can help direct future research and improve molecular diagnostics for individuals with neurological conditions.

Matching journals

The top 8 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.