Question

Performing DE analysis using edgeR for PhIP-seq data with multiple samples, some with some without replicates.

0

Entering edit mode

f_rahmdani • 0

@f_rahmdani-23913

Last seen 16 hours ago

Canada

Hi,

I am currently attempting to perform a differential expression (DE) analysis using edgeR for phip-seq data. The analysis involves comparing multiple samples with bead-only samples. some of the samples lack replicates while others have them, I have not encountered any errors. However, I am concerned whether this approach is appropriate for conducting the analysis.


for (cond in 1:length(pair_names)){
    CONTRASTS = makeContrasts(pair_names[cond], levels = design )
    contrast= CONTRASTS[CONTRASTS[,1] != 0,]
    names(contrast)
    contrast_pair = c(names(contrast))
    res_edgeR = exactTest(edgeR_obj, pair = contrast_pair )
   }

Thank you,

edgeR • 779 views

ADD COMMENT • link 21 months ago • updated 20 months ago f_rahmdani • 0

score 1 · Answer 1 · 2023-03-29

1

Entering edit mode

Gordon Smyth 52k

@gordon-smyth

Last seen 16 hours ago

WEHI, Melbourne, Australia

edgeR does not require replicates for every group, so there is no problem with that.

I can't make any comments on your code however. The code depends on variables that aren't defined and it doesn't entirely make sense. The final value res_edgeR is overwritten for each condition, so you'll only get results for the last condition. I'm not sure what you're trying to do, but I suspect there is a simpler way to achieve it.

ADD COMMENT • link 21 months ago Gordon Smyth 52k

0

Entering edit mode

Thank you so much for your valuable feedback. The code I shared is only a part of the overall script. To provide the results for each of the 48 experimental conditions vs control, I am writing the results to a file. Although as you said there may be simpler ways to achieve this:

group = as.factor(sample_meta$group)
design <- model.matrix(~ 0 + group)

pair_names = c()

for (col in 2:ncol(design)) {
  pair = paste0(colnames(design)[col], "-", colnames(design)[1])
  pair_names <- append(pair_names, pair )
}

for (cond in 1:length(pair_names)){
    CONTRASTS = makeContrasts(pair_names[cond], levels = design )
    contrast= CONTRASTS[CONTRASTS[,1] != 0,]
    names(contrast)
    contrast_pair = c(names(contrast))
    res_edgeR = exactTest(edgeR_obj, pair = contrast_pair )

    topTags_table = topTags(res_edgeR, adjust.method = "fdr", n = 38000)

    topTags_file = paste0(outPath_tables, "edgeR_DE_res_", contrast[2], ".tsv")
    write.table(topTags_table, topTags_file, sep = "\t", quote = FALSE, row.names = FALSE) 
 }

Once again, thank you for your assistance.

ADD REPLY • link 20 months ago f_rahmdani • 0