Hi all.
I'm new to R and really enjoying the cummerbund package. I have three questions regarding the graphs cummerbund produces.
1) Even though I have several biological replicates, my error bars in the expression plots for genes are huge. When looking at a specific gene, having run cuffdiff with 4+4 samples, the error bars are really long, all the way to the ground and high up. Why is this and why is there no difference when I have several replicates as opposed to none? Shouldn't the error bars decrease? How do I edit the graph so the error bars accurately can display the variance between replicates?
2) I cannot find the argument for y limits in the csVolcano plot. I have used the argument: xlimits = c(-10, 10) and it works perfectly but what is the equivalent for y? I have tried ylimits, ymin, ymax and nothing works.
3) When testing 3 conditions against each other, I get different number of differential expressed genes depending on if I use the getSig() function or counting the number of rows as they did in the Protocol paper. Which one is more correct?
gives 315 genes, while
gives 372 genes.
Thank you for any help!
I'm new to R and really enjoying the cummerbund package. I have three questions regarding the graphs cummerbund produces.
1) Even though I have several biological replicates, my error bars in the expression plots for genes are huge. When looking at a specific gene, having run cuffdiff with 4+4 samples, the error bars are really long, all the way to the ground and high up. Why is this and why is there no difference when I have several replicates as opposed to none? Shouldn't the error bars decrease? How do I edit the graph so the error bars accurately can display the variance between replicates?
2) I cannot find the argument for y limits in the csVolcano plot. I have used the argument: xlimits = c(-10, 10) and it works perfectly but what is the equivalent for y? I have tried ylimits, ymin, ymax and nothing works.
3) When testing 3 conditions against each other, I get different number of differential expressed genes depending on if I use the getSig() function or counting the number of rows as they did in the Protocol paper. Which one is more correct?
Code:
diffGeneIDs <- getSig(cuff, level="genes", alpha=0.05)
Code:
gene_diff_data <- diffData(genes(cuff)) sig_gene_data <- subset(gene_diff_data, (significant == 'yes')) nrow(sig_gene_data)
Thank you for any help!
Comment