Sampling Defective Pathways in Parkinson Disease

Abstract

Discrimination of case-control status based on gene expression differences has potential to identify novel pathways relevant to neurodegenerative diseases including Parkinson’s disease (PD). In this paper we applied two different novel algorithms to predict dysregulated pathways of gene expression across several different regions of the brain in PD and controls. The Fisher’s ratio sampler uses the Fisher’s ratio of the most discriminatory genes as prior probability distribution to sample the genetic networks and their likelihood (accuracy) was established via Leave-One-Out-Cross Validation (LOOCV). The holdout sampler finds the minimum-scale signatures corresponding to different random holdouts, establishing their likelihood using the validation dataset in each holdout. Phenotype prediction problems have by genesis a very high underdetermined character. We used both approaches to sample different lists of genes that optimally discriminate PD from controls and subsequently used gene ontology to identify pathways affected by disease. Both algorithms identified common pathways of Insulin signaling, FOXA1 Transcription Factor Network, HIF-1 Signaling, p53 Signaling and Chromatin Regulation/Acetylation. This analysis provides new therapeutic targets to treat PD.

Addressing Noise and Estimating Uncertainty in Biomedical Data through the Exploration of Chemical Space

“…In this expression, is the prior distribution for sampling the genetic signatures, and is the likelihood of the genetic signature , which depends on its predictive accuracy . These algorithms have been recently applied to perform the robust sampling of the altered pathways in different diseases: Parkinson’s, Alzheimer’s, multiple sclerosis, multiple myeloma, and triple-negative cancer [ 19 , 20 , 21 , 22 , 23 ].…”

Robust Sampling of Defective Pathways in Multiple Myeloma

“…The defective pathways are sampled via two different algorithms: Fisher’s ratio [4] and holdout [5] samplers. These algorithms have been used to unravel the altered pathways involved in the metastasis in triple negative breast cancer outperforming Bayesian networks [6]—and recently in Parkinson disease [7] —to provide new insights about the defective pathways which are involved. The problem addressed in this paper does not consist in just solving the classification problem involved in phenotype prediction, but in finding the genetic pathways that are involved in the genesis and development of this disease, which is hampered by the high degree of under-determinacy of these kind of problems.…”