Jaime Darce

Intro

using data from elsewhere compares two sets of cells - focus on regulatory t-cells wild type vs. missing one gene (PRDM-1)

what's the difference between normal T-regs and those that lack the gene?

in field, it's thought that reg cells behave differently based on particular transcription factors

because data is from another lab, it's based on a different platform and needs to be translated somewhat

Analysis 1

- Trying to upload data to genepattern, runs into errors right away attributed to the difference between the two platforms - editing a config file by hand - goes to IT guy for help some dead air on the recordings - returns with solution - goes to GP website - uploads data to multiplot, chooses to launch visualizer - Picks some settings and generates a scatterplot

? trying to get the genes that are significantly expressed by a fold change vs. coefficient of variance 3 replicates in dataset expression values with a lot of variance looking for consistent and plausible patterns (not noise or unreproducable effects) ...thus, why variance is so important

- highlighting a set of points by selecting some criteria needs to go to a rather large dialog box to do so when criteria are changed it doesn't show up immediately; must hit a "Plot" button jaime sometimes plots intermediate steps anyway - uses two different colors (red and blue) to highlight different things - looking for other files gets frustrated

? not very familiar with this data highlighted genes are overexpressed in one condition vs. another different naming conventions between the two datasets

gives up on this analysis and starts a different one.

Analysis 2

- uploading new file < i like to turn the data upside down, sideways looking for realness if you try different plots, different views, and still see something, can be more reassured that these genes are differentially expressed. do these genes fall within this signature or not? in the end it's always the same, right, you want to see the diff b/w one population and the other regular cell vs. modified foxb3 already observed the phenotype. what's the difference b/w cells that are normal vs. modified ones two different mouse lines as well: NOD vs. pb6 (?) - brings up a scatterplot with a correlation line - highlighting red and blue groups - changing plot to see if outliers are still outliers

? looking to see if outlier-ness is unique to regulatory cells. it is for blue genes, but not for red ones. so blue are more interesting (i wrote "red are more" in my notes... but that makes no sense. check video.)

- looks at volcano plot, which focuses on p-values - zooming in on the blue points? - filtering signficantly over-expressed genes

getting a signature for blues are they also in cells infected by this factor? another set of data is included now; look at cells infected w/ mt vector vs. a vector containing foxb3. taking earlier group and looking at it in another dataset. the differentiation requires more than just one factor. interactions of factors and association w/ mutation now i can go back and see what's missing w/o PRDM-1 (which was a factor in analysis 2) different between populations - how to get to the mutant you can do this sort of thing for hours!

output? a signature seeing genes uniquely expressed by T-regs that have foxb3 learning how reg cells function followup with in-depth biochemical analysis where can i go next? association of foxb3 with transcription factors? what level is the effect at? now do more specific experiments then do this same analysis again