library('data.table')
gtf = fread('/home/share/genome/dm6UCSC/dm6.refGene.gtf.gz')

print(dim(gtf))
print(gtf)

[1] 533230      9
            V1      V2          V3       V4       V5     V6     V7     V8
        <char>  <char>      <char>    <int>    <int> <char> <char> <char>
     1:   chrX refGene  transcript 11933705 11983915      .      -      .
     2:   chrX refGene        exon 11933705 11936195      .      -      .
     3:   chrX refGene        3UTR 11933705 11935993      .      -      .
     4:   chrX refGene         CDS 11935997 11936195      .      -      1
     5:   chrX refGene        exon 11936261 11936505      .      -      .
    ---                                                                  
533226:   chrY refGene        5UTR  3022552  3022602      .      +      .
533227:   chrY refGene         CDS  3022603  3023035      .      +      0
533228:   chrY refGene        exon  3023507  3024026      .      +      .
533229:   chrY refGene         CDS  3023507  3024026      .      +      2
533230:   chrY refGene start_codon  3022603  3022605      .      +      0
                                                                                                              V9
                                                                                                          <char>
     1:                                           gene_id "cac"; transcript_id "NM_001370018";  gene_name "cac";
     2: gene_id "cac"; transcript_id "NM_001370018"; exon_number "1"; exon_id "NM_001370018.1"; gene_name "cac";
     3: gene_id "cac"; transcript_id "NM_001370018"; exon_number "1"; exon_id "NM_001370018.1"; gene_name "cac";
     4: gene_id "cac"; transcript_id "NM_001370018"; exon_number "1"; exon_id "NM_001370018.1"; gene_name "cac";
     5: gene_id "cac"; transcript_id "NM_001370018"; exon_number "2"; exon_id "NM_001370018.2"; gene_name "cac";
    ---                                                                                                         
533226: gene_id "FDY"; transcript_id "NM_001370012"; exon_number "2"; exon_id "NM_001370012.2"; gene_name "FDY";
533227: gene_id "FDY"; transcript_id "NM_001370012"; exon_number "2"; exon_id "NM_001370012.2"; gene_name "FDY";
533228: gene_id "FDY"; transcript_id "NM_001370012"; exon_number "3"; exon_id "NM_001370012.3"; gene_name "FDY";
533229: gene_id "FDY"; transcript_id "NM_001370012"; exon_number "3"; exon_id "NM_001370012.3"; gene_name "FDY";
533230: gene_id "FDY"; transcript_id "NM_001370012"; exon_number "2"; exon_id "NM_001370012.2"; gene_name "FDY";

geneList = NULL
for(i in gtf$V9){
    tmp = strsplit(i,split = ';')[[1]][1]
    tmp = sub(pattern = 'gene_id \"',replace = '',x = tmp)
    tmp = sub(pattern = '\"',replace = '',x = tmp)
    geneList = c(geneList,tmp)
}

gtf$gene = geneList
gtf

res = gtf[,c(1,10)]
dim(res)
res = res[!duplicated(res)]

res

1. 533230
2. 2

table(res$V1)

                 chr2L                  chr2R                  chr3L 
                  3416                   3486                   3384 
                 chr3R                   chr4       chrUn_CP007081v1 
                  4088                    105                      2 
      chrUn_CP007120v1       chrUn_DS483646v1       chrUn_DS484313v1 
                    20                      1                      4 
      chrUn_DS484385v1       chrUn_DS484581v1       chrUn_DS485228v1 
                     4                      1                      5 
      chrUn_DS485324v1       chrUn_DS485453v1       chrUn_DS485536v1 
                     4                      5                      4 
      chrUn_DS485826v1       chrUn_DS485964v1       chrUn_DS485968v1 
                    12                      4                     12 
      chrUn_DS485991v1                   chrX chrX_DS483995v1_random 
                     7                   2589                      1 
chrX_DS484101v1_random                   chrY chrY_CP007113v1_random 
                    12                     54                      1 
chrY_DS483742v1_random chrY_DS483875v1_random chrY_DS483931v1_random 
                     1                      4                      2 
chrY_DS484142v1_random chrY_DS484909v1_random chrY_DS485423v1_random 
                     1                      1                      1 
chrY_DS485523v1_random chrY_DS485938v1_random 
                     1                      1 

res = res[grep('chrUn',res$V1,invert = TRUE),]
res = res[grep('random',res$V1,invert = TRUE),]
res

write.csv(res,file = 'gene_chr.csv')

up = fread('roxRNAi_DEG.csv')
up = up[up$status=='up',]
head(up)

library(dplyr)
table(res$V1[res$gene %in% up$names])

载入程辑包：‘dplyr’


The following objects are masked from ‘package:data.table’:

    between, first, last


The following objects are masked from ‘package:stats’:

    filter, lag


The following objects are masked from ‘package:base’:

    intersect, setdiff, setequal, union






chr2L chr2R chr3L chr3R  chr4  chrX 
  860  1027   978  1215    48   881 

table(res$V1)

chr2L chr2R chr3L chr3R  chr4  chrX  chrY 
 3416  3486  3384  4088   105  2589    54 

pct = NULL
for(i in names(table(res$V1[res$gene %in% up$names]))){
    pct0 = table(res$V1[res$gene %in% up$names])[i]/table(res$V1)[i]
    pct0 = round(100*pct0,2)
    pct = c(pct,pct0)
}
pct

chr2L

25.18

chr2R

29.46

chr3L

28.9

chr3R

29.72

chr4

45.71

chrX

34.03

library(ggplot2)

p = ggplot(data = NULL,aes(x = names(pct),y = pct,fill = names(pct)))+
    geom_bar(stat = "identity")+
    theme_classic()+
    scale_fill_manual(values = c('#2E7EBB','#EB2C2B','#37AA4A','#957B5B','#7262AC','#EABE0B'))+
    labs(x = 'Chr',y = 'percentage(%)')
ggsave(filename = 'ChrPct.png',plot = p,width = 6,height = 6,dpi = 300,bg = 'white')
p

tmp = table(res$V1[res$gene %in% up$names])
p = ggplot(data = NULL,aes(x = "",y = tmp,fill = names(tmp)))+
    geom_bar(stat = "identity")+
    coord_polar(theta = "y")+
    theme_void()+
    scale_fill_manual(values = c('#2E7EBB','#EB2C2B','#37AA4A','#957B5B','#7262AC','#EABE0B'),breaks = names(tmp),labels = paste0(names(tmp),': ',round(100*tmp/sum(tmp),2),"%"))+
    labs(fill = 'Chr')
ggsave(filename = 'degPct.png',plot = p,width = 6,height = 6,dpi = 300,bg = 'white')
p

Don't know how to automatically pick scale for object of type <table>.
Defaulting to continuous.
Don't know how to automatically pick scale for object of type <table>.
Defaulting to continuous.