High Feature-to-Sample Ratio Neurocystercosis Data Set in Gene Expression Microarray Analysis

Medical experiments with diseases that are diagnosed late result in small sample sizes. The sample size and interaction become extremely problematic when applying statistical methods for analysis since the Kolmogorov’s strong law of large numbers, Gaussian distribution, and independently identically distributed random variables are not possible assumptions. Gene expression analysis requires a data driven methodology due to microarrays generating large amounts of data from small sample sizes. This requires a Bayesian approach with dependence on the prior which must be mathematically deduced to maintain accuracy.The research dataset was provided by Dr. Fleury, at the National Institute of Neurology and Neurosurgery, Manuel Velasco Suárez in Mexico City, and was composed of 30 microarrays from nine Neurocysticercosis patients. This disease is caused by the larva form of the ingested pork tapeworm Taenia Solium. It asexually reproduces in the intestine and releases the larva into the blood stream allowing implantation in the brain and compromising the central nervous system. It symptomatically manifests through headaches, seizures and if left untreated may cause death. The microarrays were statistically interpreted in R with Bioconductor for twelve contrast regarding location of parasite and treatment of the disease. The basic analysis procedures consisted of Low level data analysis, gene classification, and machine learning. From the twelve contrasts there were significant up and down regulated genes as well as pathways for the relationship between genes. From looking at differentially expressed genes we can determine how to more effectively make medicine to optimally treat Neurocysticercosis.