This code was made to analyze the Bulk RNA Expression from two data-sets sequenced seperately via Novaseq. Previous to this analysis, both data-sets were analyzed seperately via DESeq2. After this initial analysis, a tab deliminated text file of significantly differentially expresssed gene symbols shared between the two results was saved (labeled in the results folder as "sharedGenes.txt." Because the data is currently in the process of publishing, the original expression datasets have not been uploaded and image plots that are visible have been altered and do not reflect the original expression data-sets.
The code begins by reading in the two expression datasets (saved as rangeSummerizedExperiment RDS files as output from tximeta), creating shared columns between the two datasets, and then finally combining these data sets based on shared columns into a single rangeSummerizedExperiment object (saved to a file). Dataset1 has 7 samples, all cells from the same macaque, 4/7 of which have a Mutant (Mut) genotype and 3/7 of which have a Wild type (HT) genotype. The second data set has 6 samples, 3/6 of which have a Mutant (Mut) genotype and 3/6 of which have a Wild type (HT) genotype. The second differs slightly in its labeling of "genotype" as "condition" (this is corrected to keep consistency between the labeling of the two datasets.
After combining the expression datasets and adding annotation, exploratory analysis is preformed to visualize the top 2 principle components and to observe trends in euclidean distance between genotype. After intial visualizations (using variance stabilized ie normalized expression data) methods DESeq2 and EdgeR are both used seperately to perform differential gene expression analysis, the results of which may be saved in heatmaps, volcanoplots, and files. After performing each method seperately, significant and differentially expressed genes common between the two methods are found and methods are compared.
Undirected gene set analysis was also performed on the differential expression results of EdgeR via goanna and kegga functions and can be visualized via custom built barchart functions. Pathway tool was also used breifly to visualize the differential expression of genes in pathway maps of interest obtained from kegg.
A final list of significant differentially expressed genes shared between DESeq2, EdgeR, and also the old separated method of analysis was then used as the starting point for module analysis. Eleven gene modules from this shared list were created using a custom function analogus to a pearson correlation blend. Subsets of the similarity and adjacency matrixes were plotted for better visualization and the final module results were analyzed via heatmap plots. Heatmap plots and data were saved to files and three modules were selected based on geotype clustering for further analysis and subsetting.
