phq-9

An extension to the earlier parts on this project of GSE318371 on Natural Killer T-cell Lymphoma with Epstein-Barr Viral infection in 12 healthy and 20 pathological cases in single cell RNA sequencing data with 30,960 genes and 407,006 cells in the array. We added to our other work by making the data of 30,960 genes with number of counts in all cells, number of features or genes that showed up in all cells, and percent mitochondrial DNA present in each gene, with actual gene name, link in document to a few of those data frames. Then ran T-SNE and saw a completely different clustering affect. That concludes the machine learning visualizations of unsupervised learning on this data with PCA, K-nearest neighbors, UMAP, and T-SNE. We will add more to analyze these samples from supervised perspective with the data we have gathered with some processing like our other work to see how well the top genes predict the healthy or pathology case in a 2 class model and then later get the genes associated with EBV and Lymphomas and blood cancers from the KEGG gene expression database of systemic, metabolic, and pathological gene associations to compare to the results of gene changes in our samples from the data we extracted with attached gene names to the samples comparing differential expression as fold change values from mean of healthy to mean of pathology. This numeric data has all been log normalized, scaled by 10,000, and then scaled by subtracting mean of gene across all cells not samples, and dividing by the standard deviation of that gene across all cells of the array (407,006 cells).

Demographic and Socio-e