Project description:Expression profiling by NanoString nCounter gene expression system Inflammatory response genes play an important role during acute HIV and SIV infection. Using an SIV macaque model for AIDS and CNS disease, we measured mRNA expression of 92 genes before and at multiple time points during acute SIV infection in three tissues: Peripheral blood mononuclear cells, mesenteric lymph nodes (MLN) and peripheral blood mononuclear cells (PBMC). Studying the overall changes of mRNA expressions over time or analyzing the correlation between the gene measurements and SIV RNA in plasma can result in limited interpretations. This is due to several reasons including but not limited to: 1) lack of prior information on how cells react to changes in gene expressions; and 2) biological responses typically involve many genes working together. To approach this problem, we combine multiple preprocessing methods with two multivariate analysis methods, namely principal component analysis and partial least square regression, to create a multiplexed set of 12 “judges”. Each of the judges simultaneously observes all the variables and emphasizes a unique type of gene expression response that could be significant, for example, depending on whether the cell responds to the absolute or relative size of gene expression changes. By incorporating multiple potential biological models of response, it is possible to identify genes that are consistently ranked as high “contributing” genes in different scenarios, i.e., genes that have a higher weight when we classify the data based on different classification schemes. We then use statistical analysis to verify that the consistently high-ranking genes are also statistically significant. We also investigate whether genes are tissue-specific, identify clusters of genes that co-vary together and study their correlation with regard to other gene clusters. The multiplex component analysis method introduced in this paper is a powerful tool to analyze complex gene datasets, identify significant genes, and generate testable hypotheses. In this study we inoculated 20 pig-tailed macaques with two SIV strains (SIV/17E-Fr and the swarm SIV/DeltaB670), and mock-inoculated 4 animals for control. Infected animals were sacrificed at 4 time points, as following: 6 animals sacrificed at 4 days, 6 at 7 days, 6 at 14 days and 2 at 21 days post-infection. Animals were perfused before necropsy to decrease blood contamination in the tissues. Organs were then culled, and RNA was extraction for NanosString analysis using a set of 92 probes specifically generated for macaque genes. Data were then analyzed by multiple preprocessing methods with two multivariate analysis methods. The following files contains raw data for the listed samples, respectively; GAMA Nanostring RAW MLN.xlsx - all MLN samples GAMA Nanostring RAW PBMC.xlsx - all PBMC samples GAMA Nanostring RAW Spleen 1.xls - PT01-04SP, PT71-74SP, PT141-144SP GAMA Nanostring RAW Spleen 2.xls - PT41-44SP, PT75-76SP, PT145-146SP, PT211-212SP GAMA Nanostring RAW Spleen 3.xls - PT45-46SP Please note that, in the associated publication, we describe a multiplexed component analysis to identify genes that highly contribute to inflammation during SIV acute infection. This analysis was done by compiling results generated by different types of data transformations and data normalizations (more than 12 for each tissue) and the relevant algorithms/mathematical processes will be discussed in detail. Therefore, only raw data is presented in this record (i.e. each sample data table contains raw data).

Project description:Expression profiling by NanoString nCounter gene expression system Inflammatory response genes play an important role during acute HIV and SIV infection. Using an SIV macaque model for AIDS and CNS disease, we measured mRNA expression of 92 genes before and at multiple time points during acute SIV infection in three tissues: Peripheral blood mononuclear cells, mesenteric lymph nodes (MLN) and peripheral blood mononuclear cells (PBMC). Studying the overall changes of mRNA expressions over time or analyzing the correlation between the gene measurements and SIV RNA in plasma can result in limited interpretations. This is due to several reasons including but not limited to: 1) lack of prior information on how cells react to changes in gene expressions; and 2) biological responses typically involve many genes working together. To approach this problem, we combine multiple preprocessing methods with two multivariate analysis methods, namely principal component analysis and partial least square regression, to create a multiplexed set of 12 “judges”. Each of the judges simultaneously observes all the variables and emphasizes a unique type of gene expression response that could be significant, for example, depending on whether the cell responds to the absolute or relative size of gene expression changes. By incorporating multiple potential biological models of response, it is possible to identify genes that are consistently ranked as high “contributing” genes in different scenarios, i.e., genes that have a higher weight when we classify the data based on different classification schemes. We then use statistical analysis to verify that the consistently high-ranking genes are also statistically significant. We also investigate whether genes are tissue-specific, identify clusters of genes that co-vary together and study their correlation with regard to other gene clusters. The multiplex component analysis method introduced in this paper is a powerful tool to analyze complex gene datasets, identify significant genes, and generate testable hypotheses.

Dataset Information

Macaca nemestrina

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets