Seeing negative log2FC are upregulating in RNA-seq data

I am using this code but still seeing that the negative values of the gene are upregulating.

DEGdf <- mutate(data, sig=ifelse(data$pvalue<=0.05 & data$log2FoldChange>=2 | data$pvalue>0.05 & data$log2FoldChange<2, "up", "down")) 

Please see the attached image. Any help would be much appreciated.

enter image description here

It may be better to wrap within () for each block of logical expressions and we don't need the data$ within tidyverse functions (it may work in ungrouped data, but if the data is grouped, then it takes the whole column instead of the values of the columns within the group). In the compound logical expression, the second expression with log2FoldChange < 2 may be log2FoldChange < -2

The reason is that

with(data, (pvalue > 0.05 & log2FoldChange < 2))
#[1] FALSE FALSE  TRUE  TRUE

returns TRUE for those 2 cases and thus in the ifelse, the 'yes' option i.e. 'up' got selected. If we change that to -2, it may work

library(dplyr)
data %>%
      mutate(sig = ifelse((pvalue <= 0.05 & log2FoldChange >= 2)|
          (pvalue > 0.05 & log2FoldChange < -2), "up", "down"))

-output

# log2FoldChange      pvalue  sig
#1       -3.55780 0.000081500 down
#2       -3.40030 0.002576522 down
#3       -0.76560 0.116639300 down
#4       -1.01844 0.131739900 down

data

data <- structure(list(log2FoldChange = c(-3.5578, -3.4003, -0.7656, 
-1.01844), pvalue = c(8.15e-05, 0.002576522, 0.1166393, 0.1317399
)), class = "data.frame", row.names = c(NA, -4L))