Add links to DESeq2 vignette: http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html

Add links to Mike's book: http://genomicsclass.github.io/book/pages/rnaseq_gene_level.html

attr(dds, "modelMatrix")

change all instances for consistency

Hi,
First, totally new to rnaseq analysis... figured out the command line portion (star, stringtie) and also managed deseq2 in R.
So I figured out how to analyze the differential expression in R, got actual fold change values in the command line tables!! But... I can't figure out how to 'save' the resulting expression analysis as a csv file. I did this below but I have no idea where the file is; I checked all of the folders, don't see it.

write.table(res, file="ctrlVS10um.txt", append = FALSE, sep = "\t" )

The other issue is when I tried plotMA or plotcounts, it looks like its running, but I have no idea where to find the graphs. Is it supposed to pop up in my command line window? How would I find the graphs so I can save and export them onto my local computer?

I've been googling both, but all I find are instructions on how to generate graphs or save files, nothing on how to actually save them or where they go.

Thank you.

Or we have it match Salmon

• Modify the DGE repo for consistency with changes made in last workshop
• Remove added section for resKD added in last workshop

We should be specifying the alpha value in the results() function when extracting our results. By default it tests against a alpha = 0.1, and since we select genes with padj < 0.05, we should be testing for an alpha = 0.05.

Also, if we use a lfcThreshold, we should specify that in the results() function as well. I think this is what Mike had mentioned to us previously.

• create separate contrast vectors for OE and KD
• For OE save the unshrunken FC to plot MA with and without shrinkage

code below to use log transformed values

# Extract data from object
norm_counts <- sleuth_to_matrix(de, "obs_norm", "est_counts")
log_norm_counts <- de$transform_fun_counts(norm_counts)

# Compute PCs
pc <- prcomp(t(log_norm_counts))
plot_pca <- data.frame(pc$x, summarydata)


# Plot with sample names used as data points
ggplot(plot_pca, aes(PC1, PC2)) + 
  theme_bw() +
  geom_point(aes(color=genotype)) +
  xlab('PC1') +
  ylab('PC2') +
  scale_x_continuous(expand = c(0.3,  0.3)) +
  #geom_text_repel(aes(x=PC1, y=PC2), label=name) +
  theme(plot.title = element_text(size = rel(1.5)),
        axis.title = element_text(size = rel(1.5)),
        axis.text = element_text(size = rel(1.25)))

Hi,

I believe there might be an error in the code for stripping version ids from ENSEMBL IDs because some version numbers are double digits. (lesson 9a)

If you apply current code to a file that contains any double digit ensembl version ids (including the demo salmon files):
Current code: ids.strip <- str_replace(ids, "([.][0-9])", "")
Then the second number in the ENSEMBL version ID is appended to the end to the ENSEMBL IDs which results in errors in downstream processes.
ENST00000339924.12
becomes
ENST000003399242
instead of
ENST00000339924

Probably a more efficient way to do this but I circumvented this by running two code steps to strip double digit and then single digit version ids:
ids.strip <- str_replace(ids, "([.][0-9][0-9])", "")
ids.strip <- str_replace(ids.strip, "([.][0-9])", "")

Best,
Sam

Bring in materials from Mary's BOSC lesson, specifically the table for normalization methods and associated text.

Hello,

I am new to R program, and i can not follow the step in "Overview of DGE Analysis Workflow" from: Salmon (quant.sf) to tximport. https://hbctraining.github.io/DGE_workshop_salmon/lessons/01_DGE_setup_and_overview.html

when i run the below codes:

List all directories containing data

samples <- list.files(path = "./data", full.names = T, pattern="salmon$")

Obtain a vector of all filenames including the path

files <- file.path(samples, "quant.sf")

Since all quant files have the same name it is useful to have names for each element

names(files) <- str_replace(samples, "./data/", "") %>%
str_replace(".salmon", "")

It showed" files character(0) & samples character (empty)", and i can not run tximport.

Look for the advice. Really appreciate.

I have checked the DESeq2 normalizaion result, the sum are also different between samples like FPKM/RPKM/TPM. In your idea, i think use RPGC(1x average coverage) are better.

In most paper(about cancer) when show the gene change quote from TCGA adopt the FPKM instead of DESeq2 normalization result( TCGA produce the raw count)

Hi,
I have a .txt file generated from featureCounts file generated from featureCounts. I want to use DEseq2 for Differential expression analysis. Please suggest any script to run the program.
Here is my input file looks like:
counts.txt

Thank you,

the images need to be updated, but maybe also some code?

The example PCA plots are too small ; increase the datapoints and the axis titles.
Also Batch should not be a continuous variable

This is helpful http://www.r-graph-gallery.com/38-rcolorbrewers-palettes/

In https://hbctraining.github.io/DGE_workshop/lessons/04_DGE_DESeq2_analysis.html#:~:text=The%20DESeq2%20dispersion%20estimates%20are,only%20based%20on%20their%20variance

There is a broken link to the 'count normalization' lesson.
I think it should go here: https://github.com/hbctraining/DGE_workshop/blob/master/lessons/02_DGE_count_normalization.md

https://github.com/hbctraining/DGE_workshop/blob/master/lessons/04_DGE_DESeq2_analysis.md

says

" For genes with moderate to high count values, the square root of dispersion will be equal to the coefficient of variation (Var / μ) "
Whereas it should be SD / μ.

(Judging by the formula Var = μ + α*μ^2)

use this as a template and update the link to lessons - https://github.com/hbctraining/Intro-to-ChIPseq/blob/master/README.md

It is okay if there is not an actual schedule, as long as each lesson has a time associated with it.

not sure what the equivalent is

we had a lot of people getting errors with masked functions in tidyverse. Adding a note with the package ::function notation would be helpful

or save the data to use with functional analysis