How to normalize the counts when you are looking at 3'UTR?

Question

[PK]

I want to know whether I have elevation at 3'UTR between two conditions. What I did was took all the transcripts and extracted the coverage using bedtools and I computed the RPKM for both conditions. Then I used WT over condition to calculate the log2foldchange value and I got this. To get this plot I grouped the values by transcript id and I sliced maximum 10 of exon rank. I assumed the max would the 3'UTR. so the 1 is 3'UTR. But my guide is saying is that I could be the library size. How to I normalize properly?

slice_max(rank_1, n = 10) %>% mutate(seqnum = 1:length(rank))