Why we are using filtering >0 for up and <0 for down after TopTags() to extract de genes ids?
0
0
Entering edit mode
8 months ago
Ann ▴ 40

How to extract list of DE genes after EdgeR? With summary() I get number of Up and Down expressed genes. And I would like to extract the gene IDs for Up and Down. As I read from other manuals, the common way is to use filtering $table$logFC > 0 for up and $table$logFC < 0 for down regulated after applying TopTags() (see My pipeline). If I use this way with I get the same amount of DE genes as with summary().

But I did not understand why we use this way. Why we dont use filtering $table$logFC >1.5 for up and $table$logFC <1.5 for down if we use LogFC cutoff 1.5?

My pipeline:
# filtering
keep <- filterByExpr(y, design)
y_filtered <- y[keep, , keep.lib.sizes=FALSE]

# to estimate dispersion 
y_disp_design <- estimateDisp(y_filtered, design = design, verbose=TRUE)

# Fit a quasi-likelihood negative binomial generalized log-linear model to count data
fit <- glmQLFit(y_disp_design, design, robust = TRUE)

# Conduct genewise statistical tests
LFC <- 1.5
tr.c1_2  <- glmTreat (fit, contrast = my.contrasts[, 'c1_2'], lfc = log2(LFC)) 

# Identify which genes are significantly differentially expressed
is.de.tr.c1_2 <- decideTestsDGE(tr.c1_2)

# View the  amount of Up and Down genes
summary(is.de.tr.c1_2)
Down 3191
NotSig 13930
Up 1077

# Extracts the most differentially expressed genes
TT_tr.c1_2 <- topTags(tr.c1_2, Inf)

# View TT_tr.c1_2
TT_tr.c1_2
Coefficient:  1*stage_1 -1*stage_2 
                          logFC unshrunk.logFC   logCPM       PValue          FDR
TRINITY_DN583_c0_g1   -3.254507  -3.255309e+00 6.980710 1.123231e-11 2.044055e-07
TRINITY_DN1301_c4_g1  -5.887508  -5.921260e+00 3.852718 2.272249e-11 2.067519e-07
...

# extract Up genes
up_c1_2 <- row.names(TT_tr.c1_2$table[TT_tr.c1_2$table$FDR < 0.05 & 
                                    TT_tr.c1_2$table$logFC > 0, ])

length(up_c1_2)
1077
EdgeR logfc DEG • 302 views
ADD COMMENT

Login before adding your answer.

Traffic: 2407 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6